Use-case register
Every published entry, filterable by sector, AI capability and where the system runs. Each one names the organisation behind it wherever the source does.
Independer cuts feature delivery 50% with Lakebase
Independer, the largest independent comparison platform in the Netherlands, used Databricks Lakebase to eliminate the export bottleneck between its Customer 360 data platform (built on Databricks and Unity Catalog) and the production APIs serving its website, customer service tools, and internal analytics. Because customer data had been fragmented across legacy on-premises systems, engineers previously had to write export scripts and load a separate SQL Server instance before building new features. By synchronizing gold tables directly with consuming services via Lakebase (built on PostgreSQL), Independer cut development work for new Customer 360 features by 50% and is now building GenAI agent proof-of-concepts that store state in Lakebase.
Aizon Improves Pharmaceutical Manufacturing With Artificial Intelligence Using AWS
Aizon built a GxP-compliant AI platform on AWS to help pharmaceutical manufacturers optimize biomanufacturing processes. Using AWS IoT, AWS Lambda, and Amazon SageMaker, Aizon created digital twins of bioreactors that trigger machine learning operations for process optimization. For a multinational pharmaceutical company producing blood-plasma-based drugs, Aizon built unsupervised and supervised learning models that identified two critical process variables; optimizing them improved yield by double digits, worth potentially tens of millions of dollars. AWS also reduced Aizon's compute costs by up to 80 percent.
Plenitude builds machine learning models on Databricks to forecast energy demand and renewable production
Eni-owned energy company Plenitude, which serves 10 million households and businesses across Europe, uses statistical models and machine learning on the Databricks Data + AI Platform to forecast customer energy consumption at hourly and daily granularity, forecast wind and solar generation from its renewable asset portfolio, and run customer segmentation and propensity models across 60 implemented use cases.
Aareon builds AAVA, an AI virtual assistant integrated into its German property-management ERP, using Amazon Bedrock
Aareon, a Germany-headquartered property management software provider, integrated an AI virtual assistant called AAVA into its German ERP solution Wodis Yuneo using Amazon Bedrock. AAVA lets real estate housing company customers query documents, request information through chat, automate processes, and accelerate workflows -- including scheduling appointments between tenants and craftsmen, coding invoices, analyzing images to draft damage reports, and sorting incoming email, with both text and voice support. Aareon uses various LLMs available through Bedrock, including Anthropic's Claude models and Amazon Nova models, testing and selecting models by complexity and price performance for each use case (for example, a higher token-limit model for calculating service charges). The Bedrock deployment runs in the AWS Germany (Frankfurt) Region to help meet GDPR requirements, with customer data not used to train or improve the underlying models. The first version of AAVA was built in 4 weeks and announced to customers within 6 months, developed with support from AWS Partner Reply.
Building a central data, analytics, and AI solution using AWS with Generali
Insurance and asset management group Generali built a central data, analytics and AI solution on AWS to standardize AI adoption across its more than 40 operating entities, using Amazon Redshift for the central data warehouse, AWS Glue for data-quality and transformation pipelines, Amazon SageMaker AI to automate claim settlement and predictive payout modeling, and Amazon Bedrock to power a generative AI customer-assistance agent providing instant responses. The company developed 16 flagship AI use cases spanning pricing, underwriting, claims processing and operations, deployable across operating entities via a centralized Global AI engine, while complying with GDPR and internal governance policies.
A Journey Towards Unified Data Governance at bp
bp implemented Databricks Unity Catalog to unify data governance across data previously spread across disparate systems, migrating over 270 Databricks workspaces from two separate data hub platforms and 15 standalone instances, serving 10,000+ users across Finance ERP, Customer and Products, Production and Operations, and Enterprise business entities. The migration, run in four phased quarters, delivered centralized governance, granular row/column-level access control, and automated column-level data lineage and audit logs. bp expects the consolidation to save nearly $1 million annually plus roughly $4,000 weekly by eliminating manual Overwatch job management, and the migration unlocked Delta Sharing, SQL Serverless Warehouses, and further AI features like model serving and fine-tuning.
GoGuardian: Safer schools, empowered teachers, thriving students
GoGuardian, which powers safe, focused learning for half of U.S. K-12 students, migrated its ML infrastructure to Databricks to manage billions of daily inferences for web filtering, classroom management and harm prevention while maintaining a PII-free, COPPA/FERPA-compliant data environment. Using Delta Lake, Lakeflow, Unity Catalog, MLflow and Databricks Model Serving, GoGuardian achieved up to 50% reduction in machine learning operational costs, 90% operational cost savings with its Delphi website classification model, and a 62% reduction in inappropriate device use among students. AI-driven prioritization also cut the volume of records requiring human review for high-risk content by over 95%, from 1 million to 35,000-45,000.
Generali France Rides the Wave of Microsoft 365 Copilot and Agentic AI with Microsoft
Generali France rolled out Microsoft 365 Copilot, Copilot Studio, and Azure OpenAI as part of its 'Boost 2027' strategic plan. 100% of employees have access to Copilot Chat and 3,700 use Microsoft 365 Copilot, with 70% of the workforce adopting it and generating an average of fifteen prompts per user per week. About fifty agents built with Copilot Studio address specialized cases including helpdesk operations. A 24/7 voice assistant resolves nearly 1.3 million calls (30% of requests) directly, with no human intervention. Generative AI also supports home insurance subscription, compliance beneficiary searches, and claims management, where 2.1 million operations were processed by RPA robots in 2024, including automated analysis of wall cracks during drought episodes (10% needing no human intervention) and recognition of more than 15 document types. Implementation was supported by Microsoft partner Avanade.
Banking Innovator bunq Supports Growth, Strengthens Security Using AWS
bunq, a Dutch neobank with over 11 million users across Europe, uses Amazon Bedrock for several generative AI use cases including summarizing new user data with large language models, removing the need for agents to process onboarding documents manually. Using Amazon Bedrock, bunq tripled user support process efficiency while maintaining over 90 percent accuracy. Sensitive data stays within bunq's AWS virtual private cloud, supporting GDPR and PCI DSS compliance alongside tools such as AWS CloudHSM, AWS Security Hub and AWS KMS.
Foxintelligence Enhances Ecommerce Insight Using AWS and Generative AI
Foxintelligence by NielsenIQ, the largest ecommerce measurement and consumer analytics company in Europe with around 5 million active online buyers tracked, uses Amazon SageMaker to build, train, and deploy its machine learning models and is exploring Amazon Bedrock to access foundation models, fine-tune them, and do prompt engineering, reducing time to market for generative AI innovation. As a company handling consumer personal data, Foxintelligence relies on AWS security features to comply with the EU's General Data Protection Regulation (GDPR) and other regional regulations as it expands globally, with advice from the AWS engineering team on new compliance features.
CDPHP modernizes infrastructure and improves medical data extraction with AWS AI/ML
CDPHP, a not-for-profit health plan serving 400,000 members in Upstate New York, used AWS services including Amazon Comprehend Medical, Amazon Textract, and Amazon SageMaker to automate its data processing pipeline for unstructured medical records and health data. The organization processed over seven million records during initial migration and now processes 3,000 electronic health records weekly. CDPHP achieved a 60% improvement in overall efficiency and reduced HEDIS report generation from 4-5 days (three data scientists) to two reports produced daily.
Contra Costa County District Attorney's Office Makes Unbiased Charging Decisions with ScaleCapacity Generative AI Solution on AWS
The Contra Costa County District Attorney's Office worked with AWS Partner ScaleCapacity to build a Race-Blind Charging solution to comply with California's AB 2778 mandate. The solution uses Amazon Bedrock and Amazon Textract to automatically redact race, ethnicity and other identifying details from police reports and case documents before charging decisions are made, with Amazon S3, DynamoDB, Cognito and SES supporting document storage, metadata and user access. The office processes around 17,000 cases annually, achieved compliance in six months, and can test and deploy new redaction rule changes in under a week.
Unipol accelerates insurance operations with Generative AI built by Reply
Italian insurer Unipol Group, serving customers through more than 2,000 agencies, worked with Reply (Cluster Reply and Sprint Reply) to build a governed Generative AI platform: a Foundation Layer/Unipol GenAI Stack using Azure AI Search and a Vector DB indexing over 60,000 documents; a Virtual Information Assistant chatbot with RAG over 50,000+ documents that reduces back-office ticket load and delivers responses in seconds; a network of AI agents built with Azure OpenAI, LangChain/LangGraph and UiPath that convert manual tasks into conversational interactions with real-time data validation; and an AI Guardrail governance layer (built on the LoopMind asset) providing LLM-as-a-judge monitoring, human-in-the-loop validation and AI Act compliance. Reply reports accelerated time-to-value, with new applications released in 1-3 months.
ThinkDeep's DeepBrain AI agents help automate public services for the French government
ThinkDeep AI developed DeepBrain, a multi-agent assistant platform for France's Ministry of Economy and Finance and Ministry of Defense, built on NVIDIA AI Enterprise (NIM, NeMo Retriever, NeMo Guardrails, Llama Nemotron models) running on-premises on NVIDIA DGX H100 systems to keep sensitive data within government infrastructure and comply with the EU AI Act. DeepBrain agents process millions of PDF documents, scanned images, schemas and videos for use cases including fraud detection and legal document processing. Document retrieval time fell from two days to two minutes, a typical 100-user deployment saves 1,000 work hours per month, and the Ministry of Finance has saved €2 million deploying DeepBrain at scale to 10,000 employees.
Efficiency meets data sovereignty: Lector automates document processing with AI
Lector, a platform for intelligent document processing used by more than 60 corporate customers including over 25 health insurers, migrated its productive cloud services from a major American cloud provider to OVHcloud's German data centre to meet GDPR and data-sovereignty requirements. Using OVHcloud Managed Kubernetes, S3-compatible Object Storage, and AI Training/AI Deploy/AI Endpoint services, Lector achieved over 80% time savings in document processing across more than 500 million pages processed to date, while reducing the operational burden of running Kafka and MongoDB themselves.
Boosting Innovation and Cutting Costs Through Lockheed Martin's AI Factory
Lockheed Martin centralized compute resources, MLOps tools and best practices into the Lockheed Martin AI Factory, built on an NVIDIA DGX SuperPOD reference architecture, to build and deploy trustworthy AI at scale on-premises under strict data governance requirements. Developers can now get GPU-backed environments running in minutes instead of weeks, and training times dropped from weeks to days. The AI factory processes over one billion tokens per week and now serves 7,000 engineers and developers, supporting internal chatbots and coding assistants such as Lockheed Martin Text Navigator, and has consolidated 30+ models on-premises.
TELUS builds its Fuel iX generative AI platform on Google Cloud
Canadian telecom TELUS built Fuel iX, a proprietary generative AI platform on Google Cloud, giving more than 57,000 team members secure access to 40+ AI models, including Anthropic's Claude and Google's Gemini via Vertex AI. Employees have created over 13,000 custom AI solutions and 47 large-scale gen AI products, driving more than $90 million in benefits and 500,000+ hours of time savings to date, and TELUS's engineering teams now ship code 30% faster. TELUS extended Fuel iX to power its customer-facing GenAI support tool, which became the first gen AI tool in the world internationally certified in Privacy by Design.
City of Vienna increases efficiency of public services with innovative AI solution
Since 2023, Wien Digital, the City of Vienna's IT department, has worked with Red Hat to give the city's 70,000 employees access to WienKI (ViennaAI), an AI-based virtual assistant built on Red Hat OpenShift on premise and Microsoft Azure Red Hat OpenShift. WienKI answers employees' work-related questions from a growing knowledge base of digital documentation. The city follows the Vienna Manifesto on Digital Humanism and the EU AI Act, keeping a human in the loop to verify accuracy and make final decisions.
Novartis: Accelerating Drug Development with AI-Powered Clinical Trial Transformation
Novartis partnered with AWS Professional Services and Accenture to modernize their drug development infrastructure and integrate AI across clinical trials, with the goal of reducing trial development cycles by at least six months. The initiative built a GXP-compliant data mesh platform on AWS with Databricks for processing, enabling AI use cases including protocol generation and an intelligent decision system (digital twin). Early results from the patient safety domain showed 72% query speed improvements, 60% storage cost reduction, and 160+ hours of manual work eliminated. The protocol generation use case achieved 83-87% acceleration in producing compliant protocols.
Operationalizing Analytics Data at Scale (DEICHMANN SE)
DEICHMANN SE, Europe's largest footwear retailer with roughly 4,700 stores in more than 30 countries and about 8.7 billion euros in 2024 revenue, adopted Databricks Lakebase, a managed PostgreSQL database integrated with its Databricks Lakehouse, to activate governed customer and product data for e-commerce marketing without building separate integration pipelines. Lakebase reduced a data-product publishing pipeline to 3 clicks and about 30 seconds, and is now in production across Europe feeding SAP Emarsys for e-commerce marketing.
Husqvarna Group Boosts Operational Efficiency with Azure Adaptive Cloud
Swedish manufacturer Husqvarna Group deployed Azure Arc, Azure IoT Operations, and Azure OpenAI Service to unify its fragmented public cloud, hybrid, and on-premises factory infrastructure and enable real-time data insights across its global manufacturing network. The rollout achieved 98% faster data deployment and 50% lower infrastructure imaging costs. Using Azure OpenAI Service, the company also built Factory Companion, an AI chatbot that gives factory teams instant troubleshooting guidance based on historical maintenance data and machine documentation, moving maintenance from reactive toward predictive.
Attensi Uses the Power of AI to Reduce the Time It Takes to Develop Content by up to 95%
Norwegian gamified-training software company Attensi enhanced its Creator authoring tool and simulation training platform with generative AI built on AWS services including Amazon Bedrock, Amazon SageMaker, and Amazon EKS, plus AWS Partner Hugging Face. The AI features generate dialogue, tasks and minigames, translate text, and convert text to speech, cutting content-creation time by 90-95% with Creator's AI copilots. Attensi's aggregate customer data shows game-based training drives a 19% productivity increase, 5x faster time to competence, 85% knowledge-gap closure, 2.5x revenue growth, and 75% attrition reduction.
Roche scales NVIDIA AI factories globally to accelerate drug discovery, diagnostics and manufacturing
Roche is deploying more than 3,500 NVIDIA Blackwell GPUs across hybrid cloud and on-premises environments in the U.S. and Europe, expanding an existing collaboration to turn AI and accelerated computing into a core operating capability across its pharmaceutical and diagnostics businesses. AI is central to Genentech's Lab-in-the-Loop drug discovery strategy, where nearly 90% of eligible small-molecule programs integrate AI; in one program a degrader molecule for oncology was designed 25% faster, and in another AI helped deliver a backup drug candidate in seven months instead of more than two years. Roche is also using NVIDIA Omniverse to build digital twins of manufacturing sites, including its new GLP-1 facility in North Carolina, uses NVIDIA Parabricks software to enable insights across vast datasets, is evaluating the BioNeMo platform for molecular design and chemistry innovation, and uses NVIDIA NeMo Guardrails to ensure safe and reliable healthcare-grade AI.
omni:us: A cloud AI pipeline brings order to insurance claim processing
omni:us, a Berlin-based firm spun out of parent company Qidenus Technologies, built a hybrid on-premises and cloud AI pipeline that sorts insurance claims paperwork and automates processing for European insurers. The system captures data on-premises and processes it using cloud AI, combining convolutional neural networks, NLP, optical character recognition and handwritten text recognition to convert unstructured claims documents into structured data, aiming to help insurers settle claims within minutes instead of weeks. For a handwritten car claims case, the model achieved a handwritten character error rate of 7.25%. The company uses Google Kubernetes Engine and Container Registry for its European clients and supports on-premises deployment for insurance customers.
bunq builds intelligent banking assistant Finn with Claude
bunq, Europe's second-largest neobank serving 20 million users, built a multi-agent banking assistant called Finn using Claude, selected after A/B testing showed superior multi-step reasoning for investigating failed payments, transaction disputes, and account issues. In 2024, bunq became the first bank to launch real-time speech-to-speech AI translation, letting users speak to support agents in their native language, alongside app translation into 38 languages. With Claude's advanced reasoning, bunq's automated resolution rate improved by 8%, bringing total support automation to over 80% of tickets, account opening now takes 5 minutes instead of days, and new developers using Claude Code contribute meaningful code in 16 hours instead of 40 (a 60% reduction).
GoJob enhances recruitment capabilities with Mistral
Gojob, a French HR-tech platform specializing in temporary worker staffing that processes over 2 million applications and completes approximately 60,000 recruitments annually across Europe and North America, integrated Mistral Large into its recruitment platform on Azure. The GenAI recruiting assistant conducts real-time conversations with candidates to automate pre-qualification and initial candidate screening.
Zalando enhancies customer engagement and operational efficiency with Mistral
Zalando, a European e-commerce platform, integrated Mistral models hosted on AWS Bedrock into its platform to add natural language processing and machine learning capabilities. The integration supports personalized recommendations, improved customer service and streamlined operations for the retailer's shopping experience.
Transforming the CX experience: How Cisco plans to cut renewal proposal time by 20%
Cisco's Customer Experience organization partnered with Mistral to build the AI Renewals Agent, the first solution from a strategic partnership between the two companies. Account managers previously had to manually analyze more than 50 signals across structured and unstructured customer data to draft renewal proposals. The agent uses text-to-SQL across multiple platforms, keeping customer data on premise and scaling across Cisco's own GPU cluster, with Mistral's Applied AI team collaborating on fine-tuning. Cisco estimates the agent can cut time spent creating renewal proposals and preparing for customer engagements by up to 20%.
Miroglio Group Unleashes Fashion Innovation in Italy with Generative AI on AWS
Italian fashion company Miroglio Group, which designs and distributes 10 fashion and lifestyle brands across 41 countries, worked with AWS Partner Data Reply to replace its manual product-tagging process with a generative AI-powered tagging system built on Amazon Bedrock using Anthropic's Claude 3.5 Sonnet, combined with a machine learning model trained on the company's historical images. The system analyzes product images and generates accurate technical and editorial tags and descriptions across multiple languages, reaching almost 90 percent accuracy and cutting a task that previously took one month down to about one hour.
SAP builds a Mistral-powered multilingual chatbot that cuts SBB's software-migration support queries by 80%
To support end users during Swiss Federal Railways' (SBB CFF FFS) multi-year software migration affecting 30,000 employees, SAP built a multilingual chatbot using Mistral Large hosted entirely on SAP-operated infrastructure, making the deployment 100% European and locally hosted. The chatbot uses retrieval-augmented generation to draw on FAQs, training materials and unstructured content, with glossary-based query rewriting for domain-specific abbreviations and a human-in-the-loop process that captures expert answers into the knowledge base. It integrates into existing workflows via SAP's Build Work Zone portal and Microsoft Teams. The implementation achieved an estimated 80% reduction in repetitive support queries and scaled from 1,000 pilot users to SBB's full 30,000 employees, supporting German, French and Italian natively.
Doctolib accelerates developer productivity with Claude Code
Doctolib, Europe's leading healthcare technology platform serving 420,000 health professionals and 90 million patients, rolled out Claude Code across its entire engineering team after piloting with 30 engineers. The platform team built a centralized repository of reusable prompts, custom commands, and subagents for documentation, testing, code review, and debugging. Claude Code migrated the company's entire legacy visual regression testing infrastructure in hours instead of weeks, automatically updates technical documentation on every code change, and now handles pull request reviews instantly through automated Claude-powered reviews on their main infrastructure repository.
N26 automates customer operations with Claude
N26, a fully licensed digital bank serving customers across 24 European markets, deployed Claude via AWS Bedrock in Europe to automate document-heavy customer service workflows. Since 2024, N26 has integrated Claude into more than 15 internal applications: a customer-facing virtual assistant providing instant support in five languages, automated chargeback handling that translates documentation and recommends decisions on complex cases, processing of incoming physical claim letters, and financial crime analysis where Claude synthesizes customer data and auto-drafts investigation reports. One year after implementation, N26 automated up to 70% of tasks across targeted processes and reduced manual processing by up to 50% in specific processes.
Smarter and faster decision-making with all data in one platform
Swedbank, the largest bank in Sweden and third-largest in the Nordic countries with more than 7 million customers, migrated its analytics capabilities and application services to a Databricks-driven platform on Microsoft Azure and implemented Immuta as its unified access control plane. Using a risk-based, twofold sourcing approach that kept on-premises scenarios running in parallel during migration, Swedbank centralized markets and share trading, customer analysis, and cybersecurity data on one internal platform, making AI and machine learning available in one click and enabling shorter development cycles and faster fraud/suspicious-behavior detection.
From data migration to AI innovation: AXA Switzerland's digital transformation journey with Google Cloud
AXA Switzerland, the country's leading insurer serving over 40% of Swiss companies, migrated its data lake to Google Cloud in 2021 and built a cloud-native analytics foundation using BigQuery, Vertex AI, Cloud Run and Dialogflow. Query and processing times were reduced by over 95%, from days to minutes or seconds. The insurer uses generative AI as an assistant to human employees rather than for direct customer interaction, including natural language processing to translate and summarize complex documents such as court rulings and medical records -- cutting a task that took specialists 25 minutes down to seconds. Over 100 Google Cloud projects are now live in production across all business lines.
Novo Nordisk accelerates clinical documentation and drug development with Claude
Novo Nordisk, a global pharmaceutical company headquartered in Europe, built NovoScribe, an AI-powered documentation platform using Claude Code, Amazon Bedrock and MongoDB Atlas, to automate clinical study report (CSR) generation. Claude helped cut writing times on CSRs by 90%, reducing time spent producing clinical study documentation from 10+ weeks to 10 minutes, and delivered a 95% reduction in resources needed to create device verification protocols. The platform expanded from clinical trial reports to device protocol documentation and patient materials, and an 11-person development team, including non-technical staff, now prototypes features using Claude Code.
FraportGPT Leads the Way to the Airport of the Future: Fraport Makes Employees' Daily Tasks Easier with Azure OpenAI Service
Fraport AG, operator of Frankfurt Airport, launched FraportGPT, a company-wide employee chatbot built on Azure OpenAI Service, as its first step toward automating operations ahead of an expected 30 percent workforce reduction through aging alongside 30 percent planned growth. Programmers use it to draft code statements, lawyers use it to summarize long texts, and administrative staff use it for emails and letters and even rental-pricing formulas. Fraport plans to connect FraportGPT to in-house data sources next to further increase efficiency.
Coop Reaches 30% Retention with "AskCap", a Databricks Genie Assistant Born From a Hackathon
Retail cooperative Coop turned an internal hackathon prototype into AskCap, an AI assistant built on Databricks Genie and governed through Unity Catalog, that lets employees query enterprise data in natural language directly inside Microsoft Teams. Category managers use it for point lookups such as checking market share of specific products, and it has reached a 30% retention rate among internal users interacting with it.
Trackunit Builds IrisX on Databricks, Giving Construction Customers 5x Faster Time-to-Market
Construction telematics company Trackunit unified previously siloed Finance, Engineering and RevOps data (6 million connected machines, 3 billion data points processed daily) onto the Databricks Data + AI Platform, then exposed that capability to customers through its IrisX platform. OEMs, rental and construction companies building on IrisX see 5x faster time-to-market than building their own data architecture, and Trackunit is extending the platform toward predictive maintenance using Databricks Genie and Agent Bricks.
Bynder Reduces Search Time by 75% Using Amazon Bedrock with Amazon Titan Multimodal Embeddings
Bynder, a digital asset management company serving over 4,000 companies globally and storing more than 175 million assets (18 PB of data), built visual-similarity search using Amazon Titan Multimodal Embeddings in Amazon Bedrock. The solution converts images and search queries into vectors to match by visual and contextual similarity. One Bynder customer reports that time spent searching for assets for a typical campaign task decreased by 75%, and search results return approximately 50% more relevant options on average. Bynder does not use customer data to train the underlying large language model. The company is now exploring frame-by-frame video indexing.
Swindon Borough Council makes public information accessible with an Easy Read generative AI tool built on Amazon Bedrock
Swindon Borough Council, a local authority in England, built an Easy Read solution using Amazon Bedrock to make council documents accessible to residents with learning disabilities, low literacy or cognitive impairment. The generative AI summarizes and simplifies complex text and generates matching images, presenting simplified text alongside images on the page. The council tested the tool on a complex tenancy agreement over 50 pages long and, working with a group of Experts by Experience of a Learning Disability, refined the format based on their feedback. Combined with the council's existing Translate solution, content can be produced in 75 languages. Manually adapting a five-page document into Easy Read format previously cost around £500 and took two weeks; the council can now create an accessible multi-page document for 7-10 pence per page. The council plans to make the tool open source for other councils to use.
Volkswagen Group standardises CRM on Salesforce across 10 of 52 markets, cutting case resolution 30%
Volkswagen Group, whose brands (Volkswagen, Audi, ŠKODA, SEAT/CUPRA, Volkswagen Commercial Vehicles) previously ran disconnected customer systems, built a unified CRM approach on Salesforce's Headless 360 platform, using Marketing Cloud, Sales Cloud, Service Cloud, Data 360, Tableau and MuleSoft. Standardizing the tech stack let Volkswagen roll out to 10 European markets within two years and reach 52 countries at retail level for aftersales; giving service reps a unified view of customer and vehicle history has sped case resolution by 30%. Volkswagen Group is now planning to layer in Agentforce, Salesforce's agentic AI layer, aiming to personalize interactions across the customer journey and hand off complex cases to human reps with full context, governed by the Einstein Trust Layer.
Raiffeisen Bank International Launches Its Own ChatGPT Using Azure OpenAI Service
Raiffeisen Bank International (RBI), a cooperative bank headquartered in Vienna, Austria serving 16.5 million customers in 11 markets, built 'RBI ChatGPT' using Azure OpenAI Service and Azure AI Search within Azure AI Foundry, with Azure AI Content Safety for content moderation. The tool automates repetitive tasks like documenting intelligence and summarizing legal, regulatory and banking documents, and supports line leveling in reconciliation reporting, extracting and summarizing public news articles and financial reports, drafting emails, and answering regulatory questions. RBI ChatGPT was first deployed to 2,000 users and now has an active user base of over 20,000. A feature called 'yellowGPTs' lets employees contribute their own solutions and dedicated knowledge bases built on the platform. RBI is also experimenting with tying the o1 model on Azure OpenAI Service into its application framework.
AWS Partner Adastra Supports KWS in Generative AI Adoption Using Amazon Bedrock
KWS, a global seed producer, uses Amazon Bedrock and generative AI to boost research speed and efficiency, supported by Adastra, an AWS Partner. Access to Amazon Bedrock's large language models gives KWS flexibility to prototype and test generative AI use cases without the cost of setting up its own large IT infrastructure, according to Bjørn Øst Hansen, KWS's senior research lead for knowledge discovery.
German Federal Employment Agency deploys cloud-based AI with Privatemode
The German Federal Employment Agency (Bundesagentur für Arbeit, BA), with over 110,000 employees and one of the largest public authorities in Germany, needed an internal knowledge management AI chatbot but a local AI deployment would have been very time-consuming. It tested Privatemode, the confidential AI service from Edgeless Systems, which encrypts data during storage, transmission and processing. Using the Privatemode API, the BA set up a chatbot that answers questions from internal databases of its Cloud Center of Excellence, accessing the open-source Llama 3.3 70B model and scaling without owning GPU infrastructure, while keeping user data such as AI prompts protected against access by cloud, model, or service providers.
Roche launches NVIDIA AI factory to accelerate the development of new therapeutics and diagnostics solutions
Roche deployed 2,176 NVIDIA Blackwell GPUs on premises across the United States and Europe, bringing its combined on-premise and cloud GPU infrastructure to more than 3,500 Blackwell GPUs. The AI factory is embedded across the value chain: NVIDIA BioNeMo powers Roche's Lab-in-the-Loop strategy connecting biological and chemistry experiments with AI models in R&D; NVIDIA Omniverse-powered digital twins optimize manufacturing production lines and factory designs; accelerated computing and NVIDIA Parabricks analyze diagnostics datasets and digital pathology images; and NVIDIA NeMo Guardrails support healthcare-grade conversational AI in digital health. The expansion builds on a NVIDIA collaboration that started in 2023, and Roche's U.S. subsidiary Genentech also leverages the infrastructure for drug discovery.
MediaTek Accelerates AI Development With an AI Factory
MediaTek established an on-premises AI factory powered by NVIDIA DGX SuperPOD with NVIDIA Blackwell-based systems to accelerate enterprise AI efforts, including development of its Breeze series LLMs and a 480-billion-parameter traditional-Chinese model. The AI factory processes approximately 60 billion tokens per month for inference and completes over 24,000 model-training iterations monthly, training models exceeding 480 billion parameters within one week (versus 7-billion-parameter models in a week previously). Using NVIDIA NIM and TensorRT-LLM, MediaTek achieved a 40% improvement in inference speed and 60% increase in token throughput. NVIDIA Mission Control consolidated GPU provisioning and system monitoring, while AI-assisted code completion and an AI agent for chip design documentation reduced documentation time from weeks to days. NVIDIA Riva was integrated into NVIDIA DGX Spark for agentic voice control features like internet search, calendar and messaging.
ABN AMRO Bank migrates its Anna and Abby chatbots to Microsoft Copilot Studio, increasing Dutch intent-recognition accuracy by 7%
ABN AMRO Bank, one of the largest banks in the Netherlands, migrated its customer chatbot Anna and internal employee chatbot Abby to Microsoft Copilot Studio in a six-month project delivered with Microsoft and Capgemini. The new agents use Azure AI Language conversational language understanding (CLU) for intent recognition and entity extraction, with Azure Communication Services handling voice and Azure middleware relaying messages. The Anna agent now handles more than 2 million text conversations and 1.5 million voice conversations with customers every year, covering tasks like unblocking a debit card or changing an ATM withdrawal limit, while Abby supports employees with IT helpdesk and facilities requests. By integrating Azure AI Language CLU with Copilot Studio, ABN AMRO increased its intent-recognition accuracy rate for Dutch by 7%, reduced drop-off and transfer rates, and significantly cut operational and maintenance costs.
Erste Group Bank builds a conversational AI Governance agent on Databricks, replacing ticket-based EU AI Act compliance processes
Erste Group Bank AG, one of the largest retail banks in Central and Eastern Europe, built an Agent for AI Governance on the Databricks platform, showcased at the Data + AI World Tour in Munich. The bank replaced ticket-based processes with a conversational compliance AI assistant: users can "just talk," and the system automatically structures, validates and enriches their input for AI-governance stakeholders, producing reviewer-ready packets covering EU AI Act obligations, security, data protection and architecture while maintaining an auditable trail.
Epilot reduces email processing time by 87% using Amazon Bedrock
Cologne, Germany-based Epilot, which provides an extended-relationship-management (XRM) platform for energy companies, built an AI email-summarization feature on Amazon Bedrock (Anthropic's Claude via serverless AWS Lambda/SQS architecture) that summarizes long customer email threads for its 170+ energy-sector customers. Epilot generates 55,000 AI email summaries monthly with a negligible failure rate; about 80% of users say the feature simplifies their work, and users save 87% of the time previously spent on email management. Epilot also built a 'suggested actions' feature that auto-updates customer records from email content with a human in the loop, and keeps all processed data within an EU AWS Region using Amazon Bedrock's zero-retention data policy.
IBM Consulting Advantage integrates with Microsoft Copilot, saving consultants 250,000+ hours annually
IBM integrated its IBM Consulting Advantage AI-delivery platform into Microsoft 365 apps (Word, Excel, PowerPoint, Teams, Outlook) via Microsoft Copilot, giving consultants access to IBM AI assistants and agents from within their everyday tools. The integration is already used by thousands of IBM consultants, generating the equivalent of more than 250,000 hours saved annually, an estimated $35 million in new value; one example task went from nearly three hours to 13 minutes.
Siemens Industrial Copilot expanded, adopted by thyssenkrupp
thyssenkrupp Automation Engineering integrated the Siemens Industrial Copilot for Engineering into a battery machine used for battery quality inspections on electric cars, and plans to use the genAI-powered assistant at scale engineering machines at thyssenkrupp's global locations from 2025 onwards. The Copilot assists engineers in creating TIA Portal projects, developing structured control language (SCL) code faster for programmable logic controllers, integrating code into TIA Portal, and generating machine visualization in WinCC Unified, reducing repetitive tasks like data management and sensor configuration. Separately, the Siemens Electronics Factory in Erlangen, Germany implemented the Copilot for Operations across its soldering machines, helping operators and maintenance engineers translate machine error codes into natural language and suggest solutions, reducing machine downtime and production bottlenecks. The Industrial Copilot for Operations is planned to be offered as an on-premises hardware-software bundle with the Simatic Industrial PC, powered by NVIDIA NIM microservices, so no internet connection is required and data is processed and stored on local shopfloor hardware. Natural language processing is carried out via Azure OpenAI Service on Microsoft Cloud.
Epiroc uses Azure Machine Learning to predict steel quality and cut customer returns by 30 percent
Swedish equipment manufacturer Epiroc built an ESML AI Factory on Microsoft Azure Machine Learning and Azure Data Factory to create machine learning models for its steel heat-treatment process, predicting steel density, hardness and flexibility for rock drilling tools under various conditions. Epiroc reports the predictive models reduced customer rejections and product returns by 30 percent, and the company now runs eleven analytical teams' use cases in the shared AI factory.
Jabil builds an intelligent shop floor assistant on Amazon Q Business, giving operators instant access to troubleshooting information
Jabil, a global manufacturing and supply chain company with over 100 sites and 140,000 employees, developed an intelligent shop floor assistant -- a chatbot powered by generative AI on Amazon Q Business -- with support from AWS's Generative AI Innovation Center. The assistant gives manufacturing floor operators instant access to troubleshooting and diagnostic information, drawing on relevant documents, incident tickets, and knowledge articles from multiple systems. The first iteration was built in one week, with additional data sources integrated over following weeks. Since launch, the assistant has provided employees globally with real-time access to over 1,700 policies, manufacturing specifications, and troubleshooting documents in various languages, reducing average troubleshooting time and improving diagnostic accuracy. Jabil went on to build two further Amazon Q Business-powered assistants, an Intelligent Customer Research Assistant and an Intelligent Procurement Assistant.
Ferrari Advances Generative AI for Customer Personalization and Production Efficiency
Ferrari built a car configurator on AWS using LLMs in Amazon Bedrock and Amazon Personalize, letting customers personalize their vehicle across millions of possible configurations with 3D visualization, which increased sales leads and cut configuration times by 20%. Ferrari also fine-tuned Amazon Titan, Claude 3 and Llama models in Bedrock (combined with Amazon SageMaker JumpStart) on its own documentation to power an after-sales generative AI chatbot that classifies and summarizes customer care tickets, and uses Amazon Lookout for Vision to automate quality inspection and defect detection on the assembly line. Generative AI is also used to run vehicle design simulations and text-to-image prototyping to reduce reliance on physical prototypes.
So nutzt Krones digitale Zwillinge und KI in der Produktion
Krones, working with technology-ecosystem partners Ansys (part of Synopsys), Cadfem, Microsoft Azure, Nvidia (Omniverse/OpenUSD) and SoftServe, developed 'Agentic Digital Twins' that combine physically exact real-time simulation with AI agents capable of judgment and decision-making for beverage production lines. Simulation runs that previously took three to four hours now complete in under five minutes; the AI agents test scenarios and settings, optimize workflows, and transfer the best solutions to real equipment. The digital twin was used in developing Krones' new 'Ingeniq' line concept, presented at Drinktec 2025. Compute-intensive simulations run on Microsoft Azure using Nvidia's accelerated computing platform.
Estonia rolls out Bürokratt, an AI-guided virtual assistant network for public services
Bürokratt is a network of chatbots deployed on Estonian public sector institutions' websites, letting people obtain information from institutions and use public and information services via virtual assistants. It is a state-created, AI-based digital assistant that helps institutions deliver modern, efficient, around-the-clock customer service using large language models.
Penn Advanced Research Computing Center (PARCC)
Flexential and the University of Pennsylvania's Penn Advanced Research Computing Center (PARCC) launched a phased deployment of an NVIDIA DGX SuperPOD with DGX B200 systems and NVIDIA Quantum-2 InfiniBand networking, housed at Flexential's Philadelphia-Collegeville, Pennsylvania colocation data center, to support the university's next-generation research computing infrastructure. AHEAD, a longtime provider of enterprise IT services for Penn, handled off-site rack integration and on-site cabling.
Lola - Success story at the University of Murcia
The University of Murcia deployed the chatbot Lola, built by 1MillionBot, to handle student queries during registration and university entrance exam periods. Lola has logged over 74,000 unique conversations with about 25,000 students, with a 93.4% accuracy rate. In June-July 2019, Lola answered queries from 5,525 users across 12,247 conversations and 34,423 messages, resolving 94.2% of queries, versus 28,000 inquiries handled by staff the year before implementation.
Schneider Electric fast-tracks innovation with Azure OpenAI Service
Schneider Electric bases customer-facing AI solutions on Azure OpenAI Service within Microsoft Cloud for Manufacturing. Its EcoStruxure Microgrid Advisor uses Azure OpenAI Service and Azure IoT for dynamic control of facility energy performance, EcoStruxure Resource Advisor Copilot helps customers manage energy usage, and the company is developing a PLC code generation copilot to automate programmable logic controller programming for manufacturing robots and IoT devices.
Kadaster runs Azure Local to keep sovereign control over the Netherlands' land registry data
Kadaster, the Netherlands' official land registry and mapping agency, runs Microsoft Azure Local to maintain full control over where its data lives and how it is governed, as its workloads have grown in scale and complexity. Azure Local underpins Microsoft's Sovereign Private Cloud, which now scales to thousands of servers within a single sovereign environment.
Maya is simplifying banking for diverse populations
Maya, the Philippines' leading fintech company, partnered with Databricks to modernize its data architecture into a unified lakehouse on Delta Lake, centralizing 100% of its data for real-time access and AI scalability. Maya's ML-powered underwriting tools cumulatively disbursed $2.1 billion in loans by year-end 2024. AI-driven fraud detection, combining transactional sequence embedding, network analysis and predictive modeling, achieved up to a 98% reduction in fraud losses. Hyperpersonalized, AI-driven marketing campaigns increased average revenue per user (ARPU) by 45%. Using Unity Catalog for governance and automated ETL pipelines, Maya also achieved a 28% increase in storage capacity without a rise in costs.
Mobly triples conversion with Syte's visual AI search
Mobly integrated Syte's visual AI product discovery technology into its eCommerce platform. According to Mobly co-founder and CEO Victor Noda, the integration with Syte's visual AI technology tripled the company's conversion rate.
Terray Therapeutics uses NVIDIA DGX Cloud to train generative AI foundation models for small-molecule drug discovery
Terray Therapeutics leverages NVIDIA DGX Cloud and NVIDIA Base Command Platform to train COATI, a multimodal encoder-decoder foundation model for chemistry, reducing model training time from a week to a day and improving infrastructure utilization by over 4x compared to alternate cloud services.
Accenture and Google Cloud Support Air France-KLM in Building a Generative AI Factory
Accenture and Google Cloud collaborated with Air France-KLM to build a cloud-based generative AI factory hosted on Google Cloud that supports agentic AI capabilities, letting the airline identify, build and scale high-value gen AI, agentic and machine learning use cases. Air France-KLM has deployed a private gen AI assistant and retrieval-augmented generation use cases blending language generation with internal document search to improve diagnosing and repairing aircraft damage. The gen AI factory has accelerated development cycles by over 35%, with results already seen in ground operations, engineering and maintenance, and customer service.
Pfizer saves 16,000 hours annually with generative AI search through the PACT initiative
Pfizer and AWS created the Pfizer-Amazon Collaboration Team (PACT) initiative in 2021, pursuing 14 rapid-prototyping projects applying AWS analytics, ML and generative AI to laboratory, clinical manufacturing and supply chain challenges. Using Amazon Kendra and, later, Anthropic's Claude 2.1 via Amazon Bedrock (accessed through an internal voice/chatbot platform called Vox), Pfizer estimates scientists save up to 16,000 hours of document search time annually and cut related infrastructure costs by 55%. A separate PACT prototype uses Amazon SageMaker, Amazon Lookout for Equipment and Amazon Lookout for Metrics to detect anomalies in Pfizer's continuous manufacturing (PCMM) process. Five of the initial 14 PACT projects have moved into production.
CoreLogic cuts platform costs by 5X
CoreLogic cut platform costs by 5X after migrating its on-premises RPA platform to Microsoft Power Automate, as part of a Cognizant client engagement.
Grupo Boticário increases developer productivity by 94% with Copilot
Grupo Boticário, the world's largest beauty retail franchise with over 4,000 points of sale, standardized on GitHub Enterprise Cloud to unify a fragmented developer toolchain and made GitHub Copilot available to all 1,900 developer seats. In an internal survey, 94% of developers felt more productive and 90% reported spending less effort on repetitive tasks, with the total number of lines of code accepted via Copilot exceeding 2 million in 2024.
Axpo boosts operational efficiency across the energy value chain
Axpo Group, a major player in the Swiss energy sector, built a GenAI knowledge platform on the Databricks Data + AI Platform to give employees secure, permission-aware search across internal data sources including technical standards, specifications and best practices. The architecture uses a medallion pattern (Bronze/Silver/Gold layers), Lakeflow Jobs for orchestration, RAG-based AI Search, GPT-4o for conversational responses, and Mistral OCR for scanned documents. The tool achieved a 90% approval rating and delivered 30% time savings for power users, with a 3x increase in user adoption as it expanded beyond engineering.
Schaeffler Transforms Production With Digital Twins and Physical AI on NVIDIA Omniverse
Schaeffler Group built a digital twin platform on NVIDIA Omniverse and OpenUSD, in collaboration with Wandelbots, to plan, simulate and optimize plants, machines and workflows across its 100+ global manufacturing sites. Pilot projects in e-motor assembly (China), an inverter assembly line (Germany), and Vehicle Lifetime Solutions warehouses (Europe) cut robotic task development time from hundreds of hours to half a day and reduced commissioning times for new production lines; Schaeffler plans to integrate digital twins into more than half its manufacturing sites by 2030.
CarbonTrail cuts generative AI costs by 88% for sustainable emissions intelligence using Amazon Bedrock
New Zealand-based CarbonTrail built an AI-powered emissions measurement platform and CarbonAPI on AWS, running a document-analysis pipeline on Amazon Bedrock and workloads on AWS Inferentia to process bank-scale invoice and transactional data. The platform achieves an 87% reduction in processing time and 88% lower cost than a comparable GPT-4-with-embeddings approach, and reduces low-confidence classifications by up to 40%. The Bank of New Zealand (BNZ) uses CarbonTrail's CarbonAPI to progress toward measuring emissions across its SME customers.
Frontier Communications increases new logo sales 20% and cuts fiber pricing calculations from weeks to seconds with Databricks
Frontier Communications modernized its data foundation on the Databricks Data + AI Platform, unifying data in a secure lakehouse. Frontier's B2B sales team now uses AI-powered lead prioritization, driving a 20% increase in new logo sales; the wholesale business uses GenAI to analyze over 40,000 contracts for revenue opportunities; and fiber installation pricing calculations were reduced from weeks to seconds.
Banco Ciudad builds 10 AI agents in six months with Microsoft 365 Copilot and Copilot Studio
Banco de la Ciudad de Buenos Aires (Banco Ciudad) established an AI Center of Excellence and deployed Microsoft 365 Copilot, Copilot Studio and Azure AI Services to automate tasks and build AI agents. Starting from a pilot of 1,000 Copilot licenses, the bank built over 10 AI agents in Copilot Studio within six months to automate tasks like PDF text extraction and employee training. The bank freed up 2,400 hours annually, saved USD14,500 per month via call-center sentiment analysis on 70,000+ monthly interactions, and projects a total AI impact of USD75,000 per month, with a 90% internal approval rating for deployed AI agents, achieved without headcount reduction.
WideLabs Justice Intelligence platform makes legal services more accessible for 8 million citizens in Rio Grande do Sul, Brazil
WideLabs built the Justice Intelligence platform for the Public Ministry of Rio Grande do Sul (MPRS) in Brazil, consisting of two AI agents — the TORI Investigation Assistant and the Archiving Intelligence Assistant — plus a Citizen Access Agent for the public. The platform uses NVIDIA NeMo Retriever for RAG pipelines, NVIDIA NIM microservices for deployment, NVIDIA NeMo Curator for document preprocessing, NVIDIA NeMo Customizer for domain-specific model tuning, and NVIDIA NeMo Guardrails for safety and consistency. The system went into full production in January 2025, serving more than 8 million citizens across 497 municipalities, and per MPRS deputy attorney general João Cláudio Pizzato Sidou, procedures that could last months or years can now be resolved or advanced in less than a day.
Natura unifies data governance across 200+ AI initiatives with Databricks Unity Catalog
Brazilian beauty company Natura standardized on Databricks and Unity Catalog after previously running siloed data warehouses (Redshift, SAS, BigQuery) across countries and departments. Natura now supports more than 200 initiatives on Databricks, governs 178 data products across 19 onboarded domains, and uses AI/BI Genie for natural-language data queries. CRM initiatives built on the governed data now generate 8% of Natura's revenue.
Generali France uses AI to detect property and casualty insurance fraud
Generali France extended its partnership with Shift Technology from auto insurance into property damage insurance, deploying Shift's Force solution to detect anomalies in claims in real time by combining internal claims/contracts data with third-party sources (government data, aerial imagery, geolocation). Shift's detection focuses on 107,000 of Generali's 125,000 annual P&C claims, with roughly 3,000 claims analyzed a day and a goal of raising the recognized-fraud rate from 0.7% toward the market-leading 2% benchmark.
Accelerating the Radiological Workflow With AI at University of Wisconsin–Madison
The UW-Madison Department of Radiology uses NVIDIA DGX BasePOD with the MONAI imaging framework (integrated into Flywheel's healthcare data platform) and NVIDIA FLARE for federated learning to curate imaging datasets and rapidly iterate on AI models for tasks like pediatric bone age assessment and opportunistic screening. Ten thousand abdominal CT cases that previously took six to eight months to process manually can now be processed in a day; over one million images can be processed in under a day. The tools are being deployed via containers to a 21-site global clinical trial.
Furniture.com Transforms Online Search with Databricks
Furniture.com unifies over 60 retail partners and 1.5 million SKUs on the Databricks Data + AI Platform, using Delta Lake, MLflow and Unity Catalog to run its ML lifecycle. Its Find It AI product-discovery tool uses generative AI to create a synthetic image representing shopper intent, then matches it against the product catalog for image-based search. A Collections model uses LLMs to automatically group related products, finding more than 16,000 collections across 50+ partners with no human intervention. Users who interact with Find It AI show a click-through rate 8x higher than baseline and a return rate 3.2x higher than baseline.
Empowering Marketing Teams to Drive Greater Impact at Reckitt
Reckitt built an AI system called the Insight Engine on the Databricks Data + AI Platform, unifying consumer research, product reviews and social listening signals to give marketing teams faster, actionable insights. The system uses Databricks Model Serving to deploy multiple LLMs including OpenAI's ChatGPT-4 and Google's Gemini, AI Search for retrieval-augmented generation, and Agent Bricks AI Gateway for governance. Reckitt's marketing teams achieved a 40% improvement in task execution, a 40-60% reduction in time to develop campaign concepts, 60% faster concept development, and 30% faster ad adaptation and localization, with hundreds of marketers using the AI tools across the organization by the end of 2025.
Multitudes Builds Code Review Quality Feature in 2 Months Using 3 LLMs on Amazon Bedrock
Multitudes, a New Zealand-based engineering analytics startup, used Amazon Bedrock to build a code review quality feature, testing over 10 large language models across roughly 1,000 code reviews before choosing Amazon Nova Pro for bot detection, Anthropic Claude for feedback specificity and prompt-injection detection, and Mistral for sentiment analysis, orchestrated with Amazon Elastic Container Service. The feature increased monthly active users by 44 percent within two months of launch and reduced severe misclassification rates from 20 percent to under 1 percent.
Building a clean energy future with natural language analytics
Williams, a large-scale natural gas infrastructure operator, deployed Databricks AI/BI Genie to give commercial, regulatory, accounting and technical staff natural-language, self-serve access to analytics. The team flattened 27 disparate tables into SQL models that Genie Spaces reason over, encoding internal acronyms and business logic into Genie's instructions, powered by Databricks Unity Catalog. A data request that previously took an analyst five days now completes in seconds with validated accuracy, and the weekly backlog of data requests dropped from up to ten to one or two, freeing analysts for predictive modeling and enterprise projects.
AkzoNobel builds portfolio of 60+ autonomous AI agents on RISE with SAP and Databricks foundation
AkzoNobel consolidated 46 ERP systems down to a planned 4, standardizing on RISE with SAP and harmonizing master data with Deloitte, then connected the SAP environment to Databricks to combine structured and unstructured data for AI. On this foundation, AkzoNobel's IT organization built a portfolio of more than 60 AI agents that autonomously execute tasks such as releasing credit blocks and handling HR and logistics transactions, fully autonomously carrying out more than 750,000 actions in the first nine months, with 90% of credit blocks now resolved automatically.
Humanoid's HMND 01 completes autonomous logistics at Siemens factory in Erlangen
At a Siemens blueprint autonomous electronics factory in Erlangen, Germany, Humanoid's HMND 01 wheeled humanoid robot — running the NVIDIA Jetson Thor edge AI module for on-robot compute and developed using Isaac Sim and Isaac Lab open frameworks for simulation and reinforcement learning — completed autonomous logistics operations in a first proof of concept within the production environment. Humanoid's simulation-first development compressed what typically takes up to two years of hardware development down to seven months.
TBC Bank Operationalizes Trusted Data with Lakebase
TBC Bank, the largest banking group in the Caucasus region, built a Lakehouse on Databricks and adopted Lakebase and Databricks Apps to move from on-premises SQL Server instances and month-long reporting cycles to self-service analytics and AI-driven applications, including a web-based AI chatbot and AutoML-based credit risk scoring. Credit risk model deployment fell from 14 weeks to two days, and more than 600 users regularly query governed data through Genie.
Reinventing mobile banking with ML
HSBC used Azure Databricks and Delta Lake to unify data analytics across data engineering, data science and analysts for its PayMe mobile payments app, applying NLP and machine learning for fraud detection, personalization and recommendations. Data processing for complex analytics dropped from 6 hours to 6 seconds, 14 read replica databases were consolidated into one Delta Lake, and app engagement improved 4.5x.
BMW Group Delivers Generative AI-Based Cloud Optimization Assistant (ICCA) Using AWS
BMW Group developed the In-Console Cloud Assistant (ICCA), a conversational generative AI assistant built with Amazon Bedrock, to help its 450+ DevOps teams across more than 450 AWS accounts monitor performance, identify bottlenecks, and detect optimization opportunities. Built with the AWS Generative AI Innovation Center, ICCA keeps data within the BMW Cloud Room hosted on AWS and has helped BMW scale cloud governance and reduce costs and time to market.
REMA 1000 creates a blueprint for culture-led Microsoft 365 Copilot readiness
REMA 1000, a Norwegian discount grocery brand with nearly 700 franchise stores and part of Reitan Retail, built a culture-led governance foundation using Microsoft 365 E5 to unify Teams, SharePoint and OneDrive before rolling out Microsoft 365 Copilot. By 2025, all administrative employees had completed Microsoft Teams training and exceeded the company's 70% security knowledge benchmark. Employees now use Microsoft Copilot Chat to summarize discussions and retrieve information; one regional marketing manager described using Copilot to catch up on missed Teams channel activity after a sick day in about two minutes. The rollout reduced security risk and shadow IT by centralizing collaboration and AI usage within the Microsoft 365 tenant.
Innovating banking efficiency: National Bank of Greece transforms operations with Microsoft Azure AI Document Intelligence
National Bank of Greece (NBG), which serves 6 million customers across 356 branches, built a Document AI solution on Azure AI Document Intelligence (formerly Form Recognizer) to replace a limited rule-based OCR process. Built in under four months after analyzing 50 million pages to create a training dataset, the solution now supports about 20 document categories across nearly 60 document templates (tax forms, payroll certificates, IDs, energy certificates, utility bills), processes documents at under 0.5 seconds per page, and integrates with six customer-facing digital banking applications. Accuracy improved to approximately 90%, the system is tuned to recognize the Greek language, and NBG now processes over 700,000 pages per month.
Siemens Electronics Factory Erlangen Reduces Machine Learning Deployment Time by 80% with AWS and Siemens Industrial AI on Industrial Edge
Siemens Electronics Factory Erlangen, which manufactures PCBs and controllers such as SINAMICS converters and SINUMERIK CNC controllers, used computer vision models to spot anomalies in PCB assembly, but training and retraining ML models on premises was time-intensive and constrained by GPU bottlenecks. The factory adopted AWS services together with Siemens Industrial Edge and Siemens Industrial AI, sending shopfloor images via Edge applications to Amazon S3 before training via Amazon SageMaker or AWS Lambda; training results and edge prediction results are monitored through AI Model Monitor, with AI Model Manager providing central management of models on the shopfloor. This reduced time spent on model training and retraining by 80 percent (from about 30 minutes to roughly 5 minutes for retraining and deployment), cut costs by more than 90 percent compared to on-premises data storage systems, and reduced the false call rate by over 50 percent, while continuously preventing around 4 percent of PCB assembly errors compared to around 60 percent at peaks in the past.
Cure51 Achieves 17x Speedup in Genomic Analysis Using NVIDIA DGX Cloud
Cure51, a Paris-based Sofinnova portfolio company, achieved a 17x speedup in genomic analysis and 2x cost savings by shifting workloads to NVIDIA Parabricks running on NVIDIA DGX Cloud, accessed through the DGX Cloud Lepton program for top European healthcare and life sciences venture-backed startups, which offers access to NVIDIA H100 GPU nodes and white-glove support to help scale into markets requiring sovereign, localized compute.
Berenberg's Gemini Enterprise transformation
Berenberg, Germany's oldest private bank (founded 1590, €39 billion AUM, 1,600 employees), is the first bank in Europe rolling out Gemini Enterprise bankwide with Google Cloud. Berenberg's AI journey began in 2018 with traditional ML for portfolio management, then in 2022 the bank built BegoChat, a custom AI assistant on Vertex AI, to aggregate and analyze broker reports, company filings, and news for equity portfolio managers, infused with Berenberg's proprietary investment frameworks. A similarly built tool generates the bank's daily 'Morning Mail' market briefing, with AI assistance producing content 85-90% faster with human review for quality and compliance. Berenberg later adopted NotebookLM for an 'Everyday AI' layer of document summarization, translation, and comparison, and is now rolling out Gemini Enterprise (formerly AgentSpace) across the bank starting in 2026, giving employees a marketplace of centrally managed generic and role-specific agents (Research, financial analysis, Due Diligence support).
Building the intelligence foundation for the energy transition
BKW, a Switzerland-headquartered energy and infrastructure company, needed a more scalable, integrated approach to working with complex weather and geospatial data to support decisions across energy production, grid management, and infrastructure services as renewable penetration and climate volatility rise. BKW partnered with Microsoft to combine Aurora (an AI weather foundation model from Microsoft Research AI for Science and Microsoft AI Weather), Microsoft Planetary Computer Pro with AI Weather Grid HD EU data catalogs, and Azure-native infrastructure (Azure Container Registry, Azure Container Apps, Azure Application Insights, and an MCP server) into a unified platform for AI-driven weather and geospatial intelligence, with BKW retaining governance over how insights are applied. The collaboration is described as an exploratory, foundation-building engagement rather than a production deployment, giving BKW's data scientists, domain experts, and technology teams a shared, governed platform and supporting its net-zero and sustainable energy goals.
Novo Nordisk accelerates clinical insight with custom agents on Azure
Novo Nordisk wanted to accelerate pharmaceutical R&D decision-making by helping researchers explore and validate clinical hypotheses faster using AI-assisted quantitative analysis. Working with Microsoft's AI Acceleration Studio (Forward Deployed Engineering team), Novo Nordisk built a governed reasoning agent on Azure, drawing on its FounData initiative which harmonized more than 200,000 patient-years of clinical trial data aligned to CDISC, SDTM, and ADaM industry standards. The agent lets researchers ask complex scientific questions in natural language and generate and execute code against proprietary clinical datasets, with human-in-the-loop validation by biostatisticians before outputs influence decisions. Before production rollout, the teams ran thousands of automated tests and layered evaluation systems. The solution is reducing time to insight from weeks to minutes and increasing the number of scientific questions teams can evaluate from roughly 5-10 strong ideas per quarter to 50+.
UNIQA SEE cuts health insurance claims processing times by 90% with Azure AI
UNIQA SEE's claims handlers faced fragmented documents, manual tasks, and heavy call-center reliance that slowed claims processing. Using Microsoft Azure AI alongside Microsoft Foundry, Azure Document Intelligence, and Azure OpenAI, UNIQA SEE built an end-to-end AI-driven medical referral solution that extracts, validates, and integrates referral letter data via APIs into the core claims system. The solution reduced claims processing time from 15-30 minutes to 2-3 minutes (an 85-90% reduction), lowered costs and call volumes, and freed staff for higher-value work. UNIQA SEE also introduced a self-service web and mobile app allowing customers to submit medical referral letters digitally. UNIQA SEE became the first insurer in Southeastern Europe to implement an AI-powered business process solution of this kind.
NOL UNIVERSE builds unified data lakehouse with Databricks, cutting batch computing costs by 77%
NOL UNIVERSE, a Korean travel, leisure, and culture platform operating the NOL, NOL Interpark Tours, and NOL Tickets brands, migrated approximately 2,500 Hive-based queries and its Apache Airflow/NiFi pipelines to the Databricks Platform, consolidating disparate BI tools into a single environment and giving each business domain a self-service analytics environment governed by Unity Catalog. The customer service team also used Databricks machine learning to automate categorization of CS consultation history. The migration was completed in two months, reducing time to complete batch aggregation by 66%, decreasing batch computing costs by 77%, and increasing data availability time by 27%.
Eni runs 20-year reservoir forecast in 30 minutes using Stone Ridge Technology's ECHELON 2.0 on NVIDIA GPUs
Eni and Stone Ridge Technology (SRT) developed ECHELON 2.0, the first commercial compositional reservoir simulation software designed exclusively to run on NVIDIA GPUs, deployed on Eni's HPC5 supercomputer of 1,820 Dell EMC servers with four NVIDIA V100 Tensor Core GPUs each (7,280 V100 GPUs total). In a test, ECHELON 2.0 ran an 11.2-million-cell compositional model with nine components across a 20-year forecast in 30 minutes using eight NVIDIA A100 40GB GPUs. Eni used the software for deep-water reservoir, offshore oil field, and deep-water gas field simulations, including calibrating a strategic gas field in North Africa by running hundreds of simulations quickly.
COMLINE launches GenAI-as-a-service based on HPE Private Cloud AI from German data centers
German IT service provider COMLINE SE is expanding its cloud offering with HPE Private Cloud AI, co-developed by HPE and NVIDIA, deployed within its existing GreenLake environment operated from data centers in Berlin and Frankfurt/Main. COMLINE's first generative AI project will use HPE Private Cloud AI to classify 1.6 million legal documents per day for compliance checks for a real-estate business customer, while ensuring data sovereignty from German cloud data centers. COMLINE will also use the solution to further automate its own IT operations.
GAC R&D Center sets record-low vehicle drag coefficient using NVIDIA GPU-accelerated CFD simulation
GAC R&D Center (GAC Group) deployed NVIDIA V100 SXM2 Tensor Core GPUs on its hybrid cloud HPC platform, using Altair's ultraFluidX CFD software for GPU double-precision computing, to design a new concept car. In less than six months the team completed over 200 transient CFD simulations of vehicle outflow, achieving a simulation drag coefficient of 0.147 (wind tunnel test value 0.146 at Tongji University's Shanghai Automotive Wind Tunnel Center), a new record versus the previous 0.19. Compared to conventional CFD approaches, manual modeling effort was reduced by nearly 60% and total simulation time was shortened by about 70%.
Atlante focuses on sustainability and growth using generative AI and AWS
Atlante, the fast and ultra-fast EV charging network for Southern Europe, built its platform from scratch on AWS using serverless and fully managed services. Atlante's data science team uses generative AI services including Amazon Q and Amazon Bedrock, and is able to track carbon emissions, helping the startup increase efficiency and save money while protecting the planet.
Illimity modernizes commercial banking with Databricks and adopts LLMs for customer-centric chatbots
Illimity, an Italian digital bank, built its data architecture entirely around the Databricks Data + AI Platform to process data in real time, extract actionable insights and securely share it with partners. Databricks has supported Illimity's adoption of LLMs, using vector embedding and vector databases, to inform its knowledge base and power intelligent customer-centric chatbots. Illimity is also exploring Databricks Clean Rooms to monetize its architecture using deterministic or machine learning models without compromising confidentiality.
University of Florida builds NaviGator AI platform on HiPerGator supercomputer using NVIDIA NIM and Nemotron
The University of Florida uses its HiPerGator supercomputer with NVIDIA NIM microservices and Llama NIM to power NaviGator, a suite of self-service AI tools including chatbots and a toolkit giving students, faculty, and staff access to proprietary and open-source models. The platform is a hybrid service combining on-premises HiPerGator infrastructure with AWS, Google Cloud, and Azure, built for cost predictability, data privacy, and on-premises control of sensitive research and student PII data. Key takeaways include nearly 24,000 unique users for NaviGator Chat, 1,500+ unique uses of NaviGator ToolKit (52% from university staff), and 20 LLMs available in NaviGator.
Trustpilot boosts win rates 36% and reduces support backlog by 99% with Agentforce
Trustpilot became an Agentic Enterprise with Salesforce, using Agentforce and Data 360 to automate support and deflect cases, boosting win rates by 36%, reducing support backlog by 99%, and unifying customer data, engagement signals and account history for prioritizing prospects.
AT&T reduces fraud by up to 80% with the Databricks Data + AI Platform
AT&T migrated from a legacy on-premises, rules-based fraud detection system to the Databricks Data + AI Platform, moving to a unified lakehouse architecture. AT&T now runs over 100 ML models in production that use real-time data, automatic alerts and recommendations to protect its 182 million wireless customers from fraud including robocalls, robotexts, illegal unlocks and identity theft, reducing fraud by up to 80%.
Mercedes-Benz Consulting builds AI-powered digital assistant for natural-language document search
Mercedes-Benz Consulting transformed how employees access information by building an AI-powered digital assistant that processes natural language queries. Using Amazon SageMaker and AWS Lambda, the platform significantly improved knowledge access across millions of documents.
BMW Group reduces test case creation from 10 days to hours with an AI-powered migration factory
In a session from AWS re:Invent 2025, Frank Uslaub, Vice President of Transformation Programs at BMW Group, shared how BMW built an AI-powered migration factory to modernize legacy systems at speed and scale. Partnering with AWS Transform, BMW reduced test case creation from 10 days to hours, increased test coverage by 60%, and migrated 7 applications in just 6 months — targeting a 12-month reduction in its overall transformation timeline.
Siemens increases global search speed 3x and cuts costs 70% using Amazon Nova
Siemens wanted to help its customers navigate its many business and product lines to find what they need without sifting through multiple websites. Using AWS, the company implemented generative AI to optimize its global search: users can now enter queries in natural language and receive the most relevant information in seconds. Siemens eliminated no-results searches and improved global search speed. With access to Amazon Nova models in Amazon Bedrock, responses are three times faster, giving clients across industries high-quality information while reducing costs by 70 percent.
Novo Nordisk builds AI drug discovery platform on Azure with Microsoft Research
Novo Nordisk partnered with Microsoft Research to build an AI platform on Azure AI and data stacks spanning regulatory affairs, early research, drug discovery and trial design, using Azure OpenAI Service, Azure Cosmos DB and Azure Kubernetes Service, with Power BI and Power Apps for collaboration. The platform includes a copilot for researchers, shared reasoning-chain templates, and governance/auditing of how data and models are used. The teams published early results on predictive AI models for cardiovascular disease risk detection, including an algorithm that Novo Nordisk says predicts patients' cardiovascular risk better than the best clinical standards, drawing on more than 100 years of insulin research data.
Reveleer processes 45 million pages of medical records quarterly with AI on AWS
Healthcare data and analytics company Reveleer uses Amazon Textract and Amazon Comprehend Medical to scan medical chart data and identify diagnostic conditions with high accuracy for value-based care programs, highlighting findings for medical coders so they don't have to manually review thousands of pages per patient. The Reveleer platform maintains 100% uptime with 90% of responses served in under 8 seconds, and processed over 45 million pages of medical chart data in the first quarter of 2024. Reveleer plans to add Amazon Bedrock foundation models to its platform.
Ferrari increases custom vehicle configurations 20% with Amazon Bedrock generative AI
Ferrari used Amazon Bedrock to access and benchmark multiple foundation models, delivering a 20% increase in custom vehicle configurations and boosting productivity for over 1,000 technical users via a generative AI-powered knowledge base. Ferrari also built an AWS-enabled Car Configurator for immersive virtual vehicle configuration and achieved up to 60% faster simulations for customized models.
BMW Group boosts production efficiency with NVIDIA DGX and generative AI
BMW Group uses NVIDIA DGX systems to train deep-learning-based synthetic data generation models that power SORDI (Synthetic Object Recognition Dataset for Industries), the largest open-source dataset for industrial AI with over 800,000 photorealistic images across 80 categories. DGX systems deliver an 8x boost in data scientist productivity and 4-6x improved performance over BMW's prior legacy systems; AI training for door-sill defect detection now takes under an hour using as few as five images, and the time required for employees to implement AI automation in QA tasks overall has been cut by more than two-thirds.
Argonne National Laboratory accelerates cosmic discovery with AI-powered RADAR framework
Argonne National Laboratory and university collaborators developed RADAR, a federated, privacy-enhancing framework that lets observatories coordinate gravitational-wave and radio follow-up without moving or exposing proprietary data. Running site-local AI inference on NVIDIA GPUs across the Polaris, Delta and DeltaAI supercomputers, RADAR processed over an hour (4,096 seconds) of Advanced LIGO data in under 4.5 minutes, a 5-10x speedup versus CPU workloads. The RADAR paper has been accepted for publication in The Astrophysical Journal Supplement Series.
Myriad Genetics speeds document processing with AWS GenAI Intelligent Document Processing Accelerator
Myriad Genetics partnered with the AWS Generative AI Innovation Center to replace an Amazon Textract/Comprehend pipeline with Amazon Bedrock foundation models (Nova Pro for classification, Nova Premier for extraction) using the open-source GenAI IDP Accelerator. Document classification accuracy rose from 94% to 98%, classification cost per page fell 77% (3.1 cents to 0.7 cents), and classification time fell 80% (8.5 minutes to 1.5 minutes per document). Automated key information extraction reached 90% accuracy matching the manual baseline, with a projected $132K in annual savings and 300 hours saved monthly across 9,000 prior authorizations in the Women's Health unit alone.
Subaru selects HPE Cray XD670 servers to accelerate AI development for next-generation EyeSight driver assist system
Subaru selected HPE Cray XD670 servers, powered by eight NVIDIA H200 GPUs each, to run an on-premises AI development environment at SUBARU Lab, supporting AI-driven image recognition for its next-generation EyeSight driver assist system as part of a goal to eliminate fatal traffic accidents by 2030. The energy-efficient XD670 delivers twice the performance of Subaru's existing system, using direct liquid cooling, and ranked top in MLPerf Inference v5.0 benchmarks.
Hafnia modernizes maritime operations with Databricks Lakebase
Maritime tanker operator Hafnia built a 'data supermarket' on the Databricks Data + AI Platform and adopted Lakebase, a managed Postgres database natively integrated into Databricks, to run operational applications alongside analytics. This powers Hafnia's internal DNA Port operating portal and the Marvis AI copilot. Time to deliver production apps dropped from about two months to four or five days (90%+ reduction), and a 10-person data team now supports 10 departments, 200+ vessels and 4,500+ crew.
StoryGraph scales self-hosted LLM and ML models to 300 million monthly requests
Book recommendation platform StoryGraph, run by a two-person team, scaled AI/ML infrastructure to handle 300 million monthly requests by self-hosting large language models rather than using external APIs like ChatGPT. Self-hosting was driven by data privacy and cost: using GPT-4 at their scale was calculated to cost approximately $770,000 per month, versus a fraction of that self-hosted. The company processes roughly 1 million LLM requests per day using Redis-based queuing across GPU servers and migrated its database to a 60-server YugabyteDB cluster.
KT Corporation and Microsoft launch five-year multibillion-dollar AI partnership in Korea
KT Corporation and Microsoft announced a five-year, multibillion-dollar strategic partnership to drive AI transformation for more than 650,000 businesses and 17 million consumers in Korea. The plan includes a Korea-customized GPT-4o via Azure OpenAI Service, a Korean sovereign cloud (Secure Public Cloud) built on Microsoft Cloud for Sovereignty, a new AX-specialized service company, an AI co-innovation center with Microsoft Research, and deployment of Microsoft 365 Copilot and GitHub Copilot to KT employees, with more than 19,000 employees to be trained in cloud and AI skills.
NTT DOCOMO improves LLM usage analysis efficiency by 90% with Databricks
NTT DOCOMO built an internal 'LLM value-added platform' (approximately 10,000 monthly active users, 3 million monthly calls) and used the Databricks Data + AI Platform, including Model Serving for Azure OpenAI's GPT-4o and AI/BI Genie, to automate analysis of usage logs. This replaced manual Excel/Jupyter workflows, cutting monthly analysis work from 66 hours to 6 hours, a 90% reduction, while improving data governance via Unity Catalog.
Turkish Airlines innovates with Red Hat OpenShift AI
Turkish Airlines' technology company, Turkish Technology, deployed Red Hat OpenShift AI on premise, on bare metal, with support from a team of 5 Red Hat consultants, to build a scalable, cloud-ready AI infrastructure. The platform provides container orchestration for GPU compute resources and integrates with the Dremio distributed query engine as a data ingestion layer. Operations teams can now create data science workspaces in minutes rather than hours, and deployment times for AI models have halved thanks to standardized YAML-based deployment. AI initiatives span more than 60 live models, including real-time dynamic pricing, payment fraud detection, ground-time predictions, tail assignments, on-time performance predictions, and generative AI use cases for employee experience and productivity. More than 200 Turkish Airlines employees now work on AI-based development as citizen data scientists. The airline's Head of Data and AI, Emre Yavuz, said AI projects are targeted to create over USD 100 million in financial impact by boosting revenue, decreasing operational costs, and increasing efficiency.
Amphitrite Rides AI Wave to Boost Maritime Shipping, Ocean Cleanup With Real-Time Weather Prediction and Simulation
France-based startup Amphitrite fuses satellite data and AI to simulate and predict oceanic currents and weather, using the NVIDIA AI and Earth-2 platforms. Its fine-tuned, three-kilometer-scale AI models (dubbed ORCAst), trained on NVIDIA GPUs, analyze ocean current, wave, and wind data to help ships optimize routes and reduce fuel consumption and carbon emissions. Amphitrite trains and runs its AI models using NVIDIA H100 GPUs on premises and in the cloud, building on the FourCastNet model from Earth-2. Fusing AI and satellite imagery, Amphitrite can improve the accuracy of global ocean current analyses by up to 2x compared with traditional methods. A case study along the Mediterranean Sea found the NVIDIA-powered Amphitrite fine-scale routing solution helped reduce one shipping line's carbon emissions by 10%. Shipping and oceanographic companies using Amphitrite's solutions include CMA-CGM, Genavir, Louis Dreyfus Armateurs, and Orange Marine; the company also works with an NGO to track and remove plastic pollution in the Pacific Ocean.
Vodafone applies Microsoft's AI-powered blueprint for autonomous network operations
Vodafone is working with Microsoft to apply a proven AI-powered blueprint for autonomous network operations across transport infrastructure and field-force management, combining Vodafone's network expertise with Microsoft Foundry and the Network Operations Agent (NOA) framework. The blueprint is built on Microsoft's own experience running autonomous agents across its global Azure transport network, where AI continuously monitors performance, identifies root causes, and autonomously manages more than 65% of fibre-break field dispatches, improving time to repair by up to 25% and accelerating root-cause analysis by 80%.
Novo Nordisk Scales to 2,500+ Use Cases with Secure Generative AI Using Amazon Bedrock
Novo Nordisk built a self-service generative AI platform on AWS, using Amazon Bedrock's foundation models, so employees could build and customize chatbots for nonregulated business use cases without needing to develop applications or maintain infrastructure themselves. The company worked with AWS Partner Cloud2 Oy (previously KeyCore) to validate the architecture for security and scalability. More than 25,000 Novo Nordisk employees have used the platform to create chatbots for over 2,500 unique use cases, such as retrieving information, drafting documents, or acting as a virtual colleague or critic. The company's general-purpose chatbot is used by more than 1,000 employees and processes over 26,000 prompts a month; its largest use case was trained on 140,000 documents. Each use case costs around $10 per month to run on AWS, using serverless services including Amazon DynamoDB and AWS Lambda. Building a chatbot now takes days rather than months, and some tasks that took a full day can be completed in as little as 10 minutes.
Speed, Scale, and Smarter Care: RaySearch Delivers AI-Powered Oncology Breakthroughs
RaySearch Laboratories uses NVIDIA CUDA and TensorRT technologies across its RayStation radiotherapy planning system to accelerate image registration, dose calculations, and AI-driven deep learning segmentation for cancer treatment planning. GPU acceleration enables dose calculations up to 20x faster and full plan optimization in minutes. Iridium Network, Belgium's largest radiotherapy provider, reduced time spent on image segmentation and treatment planning by 44% and 47% respectively after implementing AI-driven automation. In 2024, over 200,000 patients benefited from automated image segmentation using this technology.
ElevenLabs: Scaling Voice AI with GPU-Accelerated Infrastructure
ElevenLabs runs its voice AI platform for text-to-speech, voice cloning and speech-to-speech translation on Google Kubernetes Engine (GKE) with NVIDIA GPUs, using GPU optimization strategies such as Multi-Instance GPUs and GPU time sharing to improve utilization and reduce costs. The platform generates 600 hours of audio for every hour of real time, supports 29 languages (expanding to 40), and is reportedly used by 41% of Fortune 500 companies, powering products including AI Dubbing Studio, long-form audio for audiobooks and podcasts, article narration for outlets like The New York Times, and real-time conversational AI assistants.
Building the foundation for the 5G revolution: Digital Nasional Berhad
Digital Nasional Berhad (DNB), Malaysia's government-established 5G network operator, used the Databricks Data + AI Platform to build a cost-effective, high-performance data platform to process 5G network data at scale, achieving 82%+ populated-area coverage in two years. Using Delta Lake with a medallion (Bronze/Silver/Gold) architecture, Databricks Workflows for pipeline orchestration, serverless compute, Notebooks for team collaboration, Unity Catalog for governance, and MLflow for ML lifecycle management, DNB achieved 70% cost optimization and a 60-70% increase in data pipeline performance, and has begun building AI-powered chatbots giving network engineers real-time access to performance insights with geospatial visualizations.
Checkr: Streamlining Background Check Classification with Fine-tuned Small Language Models
Checkr automated adjudication of background checks using a tuned logistic regression model for 98% of cases, but the remaining 2% required classifying noisy records into 230 categories. After testing GPT-4 (80-82% accuracy on hard cases, ~15s latency, ~$12k cost) and GPT-4 with RAG, Checkr fine-tuned Llama-2-7b (85% accuracy) and then, in production on Predibase, Llama-3-8b-instruct, achieving 90% accuracy on the hardest 2% of cases with 0.15-second response times (30x faster than GPT-4) and a 5x cost reduction, using LoRA/LoRAX for parameter-efficient fine-tuning and multi-adapter serving.
Working toward zero-defect production: AI in Manufacturing at Bosch
Bosch developed an in-house AI Analytics Solution now in use on over 1,400 production lines worldwide, analyzing data from highly automated, sensor-connected manufacturing to find causes of quality problems in real time. At the Bamberg, Germany plant, the platform reads and analyzes about 1 billion lines of sensor and quality data per day, processing roughly 1 million data messages every 24 hours from assembly and test lines, to flag anomalies and support zero-defect production. Bosch is working to deploy the platform across all 240 of its manufacturing sites.
Using Amazon Bedrock Agents to Accelerate Decisions Across Drug Development at AstraZeneca
AstraZeneca built Development Assistant, a multi-agent AI tool using Amazon Bedrock Agents, that lets clinical, regulatory, safety and quality teams ask natural language questions and get insights from structured and unstructured data in seconds via text-to-SQL generation and retrieval-augmented generation. A supervisor agent routes queries to specialized subagents (terminology, clinical, regulatory, database). Built on AstraZeneca's Drug Development Data Platform (3DP), the tool moved from concept to production in 6 months, including cybersecurity and AI governance checks, and is scaling to over 1,000 users.
University of Michigan Offers Several AI Services to Staff and Students
In August 2023 the University of Michigan launched UM-GPT, becoming (per the source) the first higher education institution to offer AI services at scale, providing secure access to commercially available large language models through the university's private Microsoft Azure environment. Michigan has since launched U-M Maizey, a codeless platform for building custom AI experiences from uploaded data sets using retrieval-augmented generation, with more than 3,500 Maizey instances in production across academic and operational departments, and Go Blue, a mobile AI assistant for campus life. To avoid vendor lock-in, Michigan runs AI workloads across multiple environments including on-premises infrastructure, letting users pick their LLM via a dropdown.
Toyota shifts into overdrive: Developing an AI platform for enhanced manufacturing efficiency
Toyota's Production Digital Transformation Office built an in-house AI Platform on a Google Cloud hybrid architecture that lets factory floor employees create machine learning models without AI expertise, automating manual, labor-intensive tasks. Built by a six-person team in about 1.5 years, the platform runs at all 10 Toyota car and unit manufacturing factories; the number of models created rose from 8,000 in 2023 to 10,000 in 2024, saving over 10,000 man-hours per year, with nearly 1,200 active users and 400+ trained annually.
E.ON Powers Better Decision-Making with Databricks Apps
E.ON, one of Europe's largest energy companies, used Databricks Apps to deploy internal BI applications in one to two minutes. E.ON is working on analyzing 500-600 public ChargePoint operator contracts using a GenAI-powered chatbot that extracts contract details such as pricing and tariffs via natural language queries. Interactive dashboards built with Databricks Apps give E.ON's health and safety team real-time incident data and insights, including maps of incident locations and weather conditions, to identify patterns and mitigate risk in high-voltage field work. 1,500 employees now rely on Databricks for BI dashboards and insights.
Team Liquid uses SAP Business AI to optimize esports draft strategy
Esports organization Team Liquid worked with SAP to build an AI draft bot using SAP AI Core to train a generative AI model on data from more than 6,000 professional League of Legends matches and 1.6 million amateur matches, stored in SAP HANA Cloud and running on SAP Business Technology Platform. The tool visualizes win-probability predictions after each pick and ban during the champion draft process, replacing manual analysis and saving Team Liquid approximately 10,000 hours per year.
NewDay lifts generative AI agent-assist accuracy from 60% to over 90%
NewDay, whose contact centre handles 2.5 million calls a year, built NewAssist, a generative AI assistant on Amazon Bedrock using Retrieval Augmented Generation, out of an internal hackathon. Through iterative experiments — including a custom parser for its knowledge base and injecting internal acronyms into prompts — NewDay raised NewAssist's accuracy from below 60% to over 90%, cut the average time to retrieve an answer from 90 seconds to 4 seconds, and rolled it out to over 150 agents, running on serverless AWS infrastructure for under $400 a month.
Banco Bradesco Streamlines Customer Service with Azure AI Foundry, Apps, and Databases
Banco Bradesco, co-developing with Microsoft and Avanade, built Bridge, a multi-agent, technology-agnostic generative AI platform based on Azure OpenAI in Foundry Models, with governance standardized through Azure Red Hat OpenShift and Azure API Management for deployment and scaling. Bridge powers BIA for Customers and BIA Corporate virtual assistants across phone, chat and WhatsApp for roughly 74 million customers. Results include an 83% resolution rate for digital customer service, 80% for employee queries, 89% request retention, a more than 30% reduction in technology costs, product launches up to 10x faster, an 8x productivity increase in managerial efficiency, and a 65% efficiency gain in audit planning. Azure Cosmos DB processes more than two million requests and two billion inference tokens daily. The bank runs 10 large language models and at least 20 use cases in production, including MentorIA, which automates analysis of more than 18,000 collections calls per day.
SPAR ICS Enhances Shopping and Optimizes Supply Chain with Cloud for Retail
SPAR ICS, the independent IT unit of SPAR Austria Group, used Microsoft Cloud for Retail to build the SPAR shopping app and created an award-winning AI-enabled demand forecasting system with Azure Synapse Analytics, now moving to Microsoft Fabric. The new system helps SPAR Austria Group optimize its supply chain, achieving 90 percent inventory prediction accuracy. Future plans include using Copilot for Microsoft 365 to enhance employee productivity.
RTLZWEI Hones Competitive Edge with Red Hat OpenShift AI
RTLZWEI, a German broadcaster and digital media company, modernized its on-premise core system BOBY by deploying Red Hat OpenShift Platform Plus and Red Hat OpenShift AI, with guidance from Red Hat Consulting and Red Hat Training. A proof of concept integrated NVIDIA H100 GPUs with Red Hat OpenShift AI, using Whisper on vLLM to speed up video transcription for subtitles, dubbing and accessibility. By fine-tuning the transcription model with its own data, RTLZWEI reduced the word error rate by 33%. The platform also hosts LLMs for data science experiments, forecasting and fine-tuning, running on-premise to meet data sovereignty and cybersecurity requirements.
Zapier Powers Real-Time Customer Experiences with Databricks
Zapier unified its data on the Databricks Platform to enable self-serve, AI-powered analytics across product, marketing, sales and support. Using the Databricks MCP connector built on Unity Catalog, AI agents get governed real-time access to Zapier's data; Vector Search handles over 600k requests per day for semantic search and RAG-based knowledge retrieval; MLflow and Model Serving operationalize ML workflows. Dashboard creation time dropped from 2-3 days to a few hours, and the share of data queries executed via AI-assisted workflows grew from 0% to roughly 75% within 6-9 months.
Figma trains AI Search models in 5 months using Amazon SageMaker AI
Design platform Figma built its AI infrastructure on Amazon SageMaker AI, using Amazon EMR with Apache Spark to process billions of design elements and the custom FigmaStep framework on SageMaker Pipelines to orchestrate training. Figma trained and deployed the models behind its AI Search feature within five months, launching it at Config 2024, and ran more than 10,000 SageMaker AI training jobs in 2025 to support features including the AI prompt-to-app tool Figma Make.
TK Elevator: Moving Beyond with Data + AI Driven Service Intelligence
TK Elevator (TKE), one of the world's largest providers of elevators, escalators and vertical transportation systems, adopted the Databricks Platform to unify and process more than 500 million IoT telemetry events daily from connected elevators across 30+ countries, moving from reactive to predictive maintenance. TKE's MAX platform generates predictive insights that feed actionable guidance to field technicians via its Digital Operations Center (DOC), which generates 30,000-40,000 tickets annually and has reduced technician troubleshooting time on call-backs and unplanned shutdowns by up to 20%. The platform scaled from 3 to over 40 markets (10x expansion), cut data pipeline delivery from months to weeks, and reduced data sharing time with internal/external consumers from three months to three days (90% reduction). TKE uses Delta Lake, Delta Sharing, Lakeflow, Unity Catalog, Databricks Assistant, MLflow, AI Gateway, Vector Search and Model Serving, and is exploring agentic AI to combine technical, service and commercial context in real time.
Lotte Hotels & Resorts builds a hyper-automated work culture with Microsoft Power Platform
Lotte Hotels & Resorts deployed Microsoft Power Platform, including Power Automate and Power Apps, with partners Lotte Data Communication and DEX Consulting, letting non-developer 'citizen developer' employees automate tasks after an eight-week training program. RPA applied to database refinement allowed over 50,000 pieces of data to be processed automatically that previously took 25 people nearly a month, and reservation automation across 17 hotel properties reduced 10,000 work hours per year.
How Mapfre Insurance modernized fraud claims with Amazon EMR Serverless
Mapfre Insurance, the number one auto and home insurer in Massachusetts, modernized fraud detection by combining graph-based features from Neo4j with machine learning models deployed on AWS, using Amazon EMR Serverless, Apache Iceberg tables on Amazon S3, AWS Glue Data Catalog, AWS Lake Formation, and Amazon MWAA for orchestration. Fraud predictions integrate directly with Guidewire Claims via AWS Lambda, automatically creating claim activities showing the top model drivers for adjusters. The initiative, covering Massachusetts Auto insurance and later expanded to Home insurance, has delivered more than $5 million in Net Present Value, with detection accuracy improved 50-135 percent compared to baseline methods.
Guthrie Clinic cuts patient falls with injuries by 70% using AI on Dell infrastructure
The Guthrie Clinic, serving rural communities across 11,000 square miles of New York and Pennsylvania, built an IT foundation on the Dell AI Factory with NVIDIA, including Dell PowerEdge servers and Dell AI PCs. Using the Artisight AI platform running on Dell PowerEdge servers with NVIDIA GPUs to monitor patient movement and deliver fall-risk assessments, Guthrie reduced patient falls with injuries by nearly 70%, cut ICU stays by half a day, shortened emergency room time by more than 80 minutes, and achieved $7 million in operational cost savings in 2023.
Foxconn deploys NVIDIA-powered agentic and physical AI across Taiwan's medical centers under Healthy Taiwan initiative
Foxconn's CoDoctor AI platform, with new AI agents including an ECG AI Agent, Corovia AI Agent and Endovia AI Agent, and Nurabot nursing collaborative robots powered by NVIDIA NemoClaw, have moved from pilot programs into clinical operations across Taiwan's major medical centers. This is part of the government-backed $1.5 billion Healthy Taiwan initiative, with a majority of Taiwan's medical centers already using AI for over 14 million patient encounters annually.
FactSet cuts code-generation response time 70% with a standardized Databricks LLMOps framework
FactSet, a financial data and analytics provider, standardized its GenAI development on Databricks Mosaic AI and managed MLflow after fragmented tooling across teams caused collaboration and governance problems. For its FactSet Mercury code-generation feature, FactSet fine-tuned meta-llama-3-70b and Databricks DBRX models, reducing average response latency by more than 70%. Its Text-to-Formula project reduced end-to-end latency by about 60% using a compound AI architecture with fine-tuned open-source models.
Nasdaq improves generative AI response times 30% with NVIDIA NeMo Retriever and NIM
Nasdaq built a generative AI platform for internal chatbots and search, running NVIDIA NeMo Retriever and NVIDIA NIM microservices as part of NVIDIA AI Enterprise. After running four global hackathons to prototype AI applications, Nasdaq deployed self-hosted GPU embeddings on NVIDIA L40 GPUs, achieving 30% faster response times and 30% improved chatbot accuracy while reducing embedding costs.
Techcombank: Ushering Personalized Banking for Millions of Customers
Techcombank, Vietnam's largest private financial institution with 315 branches serving over 15.3 million customers, adopted the Databricks Data + AI Platform to unify data from disparate on-premises databases and a legacy data warehouse. Central to the initiative is 'Customer Brain,' a customer-360 tool centralizing customer data for targeted marketing, and the Lead Allocation Curated Engine (LACE), which uses AI insights to prioritize and assign sales leads. Techcombank built machine learning models for fraud detection and credit risk management, using an enterprise feature store with over 7,500 features and more than 100 ML risk models to manage a credit lifecycle handling a twentyfold increase in retail credit applications. MLOps on Databricks with MLflow cut model implementation from months to weeks. The bank is developing an internal RAG-based chatbot, Smartie, built on Databricks AI Search, and piloting AI/BI Genie for natural-language data queries. The platform has over 1,000 active users bank-wide.
Supermetrics: Helping Marketers Redefine Efficiency with AI-Powered Data Analysis
Supermetrics, a Finland-based marketing intelligence platform serving 15,000+ customers across 132 countries, built an AI agent on Google Cloud using Vertex AI Agent Builder and the Agent Development Kit (ADK) that autonomously manages data connections, fixes pipeline errors, and analyzes campaign performance in real time, suggesting new creative options using Imagen. The agent automates the weekly marketing reporting cycle that previously took performance marketers up to four hours, reclaiming over 15 hours per month per marketer for strategy and creative testing. The system uses a central AI agent that interprets natural language requests and delegates tasks to sub-agents, and stores 'core memories' of user preferences for personalized context.
TMAP Mobility Powers Smarter, Safer Travel
TMAP Mobility, Korea's leading mobile navigation service with over 25 million subscribers and adoption in more than 95% of domestic vehicles, adopted the Databricks Platform to address data access and governance roadblocks that stalled AI-driven personalization efforts. Using Genie Space, non-data specialists can now explore and analyze data with natural language, and Unity Catalog provides granular permission management and column masking. As of September 2025, the analytics team receives 68% fewer simple data extraction requests, and approximately 48% of all employees can directly explore and utilize data. TMAP processes more than 7.3 billion mobility data points annually and used the Recommendation Engines Solution Accelerator to prototype and deploy scalable recommendation models faster.
Schaeffler Group: Transforming Production With Digital Twins and Physical AI
Schaeffler Group built a digital twin platform using NVIDIA Omniverse and OpenUSD to plan, simulate and optimize plants, machines and workflows across its global manufacturing operations, in collaboration with NVIDIA and robotics partner Wandelbots. The platform reduced commissioning times through virtual commissioning and facility layout planning, and cut robotic task development time from hundreds of hours to half a day using AI-driven simulation and training. At its Germany inverter assembly line, Schaeffler achieved reliable robotics performance after 24 hours of training and, with Wandelbots' virtual controller and reinforcement learning, refined robotic movements to within 5-6 centimeters for gearbox assembly tasks. Pilot projects are underway in China and Germany, including e-motor assembly and Vehicle Lifetime Solutions warehouses, with plans to integrate digital twins into more than half of Schaeffler's 100+ global manufacturing sites by 2030.
ArcelorMittal Enhances Steel Production Through Digital Innovation
ArcelorMittal partnered with IBM Consulting and Infosys to modernize operations through AI, cloud technology and SAP S/4HANA migration. At ArcelorMittal Eisenhüttenstadt, machine learning was introduced to predict and prevent surface defects on automotive steel sheets. At the Hamburg wire rod plant, AI optimized the trimming process by analyzing historical production data to determine optimal cutting points, reducing trim scrap by 20% and contributing to energy savings and lower CO2 emissions. ArcelorMittal also deployed a bio-inspired Ant Colony Optimization algorithm to calculate optimal production schedules, reducing downtime and material waste. At AM/NS India, IBM Consulting used IBM Rapid Move for SAP S/4HANA to migrate data and applications from outdated platforms to a single SAP instance across locations in Dubai, Indonesia and India.
Rabobank Streamlines and Personalizes Customer Support with AI Powered Virtual Agents
Rabobank, a multinational banking institution headquartered in the Netherlands, reconstructed its virtual agent solution using Microsoft Power Virtual Agents, replacing a disconnected on-premises IVR and chatbot setup with one conversational platform for voice and chat. Virtual agents now complete about 40 to 50 percent of the business contacts they start without escalation to a human agent, meaning about 60,000 to 80,000 calls per month are handled entirely through automation. Online chat as first point of contact rose from about 20 percent to 25 percent of customers between early and May 2022. Reporting is unified via Microsoft Power BI. Implementation began December 2021 with the first virtual agent deployed to production in April 2022, with help from Microsoft partners Avanade and Accenture.
How Thomson Reuters Turbocharged .NET Modernization with AWS Transform
Thomson Reuters used AWS Transform, an agentic AI tool for modernizing .NET applications at scale, to replace manual upgrades of legacy .NET Framework applications. AI-powered agents handled asset discovery, codebase analysis, planning, code refactoring and execution. The company modernized 1.5 million lines of code per month, a 4x boost in velocity, moved from Windows to Linux for 30% lower costs, cut application transformation time from months to a two-week sprint, and reduced technical debt by 50% by moving from .NET Framework to cross-platform .NET. AWS Transform also uncovered and fixed security vulnerabilities caused by unsupported language versions and platforms.
ABN AMRO Bank migrates to Microsoft Copilot Studio to support over 3.5 million customer conversations
ABN AMRO Bank, the third-largest bank in the Netherlands, transitioned its customer-facing 'Anna' chatbot and employee-facing 'Abby' chatbot to Microsoft Copilot Studio after a competitive RFP process, completing migration in 6 months with partner Capgemini. The Anna agent now supports over 2 million text conversations and 1.5 million voice conversations per year, and integrating Azure AI Conversational Language Understanding with Copilot Studio increased intent-recognition accuracy by 7% for Dutch.
Deutsche Telekom and NVIDIA launch sovereign Industrial AI Cloud for German manufacturers
Deutsche Telekom and NVIDIA unveiled the Industrial AI Cloud, a sovereign, enterprise-grade AI platform built in German data centers with up to 10,000 NVIDIA GPUs (including DGX B200 systems and RTX PRO Servers), NVIDIA AI Enterprise and NVIDIA Omniverse, going live in early 2026. Siemens will use the platform to accelerate industrial AI adoption; Mercedes-Benz and BMW will run AI-driven digital twin simulations for vehicle development; Agile Robots will train and validate robotic foundation models; and Wandelbots NOVA will run digital-twin-based testing and deployment on factory floors.
Colas uses AI on unified Databricks data platform to assess road maintenance needs from video footage
Colas Group worked with Databricks and OpenValue to unify data from hundreds of subsidiaries across 50 countries onto a single cloud platform combining Databricks, Microsoft Azure and open source technologies. Beyond helping reduce energy consumption and optimize industrial equipment efficiency, Colas is piloting a tool that uses AI to assess video footage of roads and determine maintenance requirements.
BMW uses NVIDIA Omniverse Enterprise and Isaac Sim to build digital twins of its factories
BMW, which produces 2.5 million cars a year with 99 percent of them customized before purchase, uses NVIDIA Omniverse Enterprise to simulate its factories in real time as a photorealistic digital twin, connecting applications like Bentley Microstation and Autodesk Revit. BMW also employs NVIDIA Isaac Sim to train delivery robots using synthetic data generation with domain randomization, and NVIDIA Fleet Command to orchestrate robots and machines across its production network.
Zoho Corporation scales agentic AI adoption on Dell AI Factory with NVIDIA
Zoho uses the Dell AI Factory with NVIDIA, featuring Dell PowerEdge XE-Series servers with NVIDIA accelerated computing, NVIDIA NeMo and NVIDIA Quantum InfiniBand switches, to train and deploy AI solutions with Dell ProSupport providing infrastructure management. Zoho delivers AI solutions such as conversational AI, Zia Agents, Agent Studio and Zia LLM to 130 million+ users across 150+ markets.
LSEG unifies 30 data systems, reducing product development time with Fabric
LSEG, a trusted data partner to 44,000 customers in more than 170 countries, partnered with Microsoft to build a unified data platform on Microsoft Fabric, consolidating 30 systems, 1,200 datasets and 33 petabytes of data. Using Apache Spark on Fabric, Apache Airflow in Fabric, Microsoft Purview and OneLake, LSEG reduced product development timelines for new data products from years to months; onboarding data and creating a new product now happens in months, accelerating launches such as its ESG and Fundamentals data products. LSEG processes around 80,000 files daily on Spark on Fabric, consuming approximately 280,000 capacity units per day, with month-on-month consumption growing more than 50%. The company is also building AI-readiness by hosting Model Context Protocol (MCP)-powered data in OneLake and enabling financial professionals to build custom AI agents via Copilot Studio.
AIBots: GovTech Singapore's platform for public officers to build RAG-powered AI chatbots
GovTech Singapore built AIBots, a platform that lets Singapore Public Officers create, customise, and deploy Retrieval-Augmented Generation (RAG) generative AI chatbots in under 15 minutes, enhancing responses by integrating internal agency knowledge bases (including content classified up to Restricted/Sensitive Normal) alongside publicly available information. Officers have used it for talking-points and speech generation, Q&A chatbots on internal HR, travel and onboarding processes, and survey verbatim analysis. Between August 2024 and February 2025, AIBots reached 40,000 users across 115 government agencies, with 12,000 bots created and over 1 million messages sent.
How Getnet improved CSAT to over 85% in under a year
Getnet, a global digital and physical payment solutions provider and part of Santander's PagoNxt payment hub, worked with partner A5 Solutions to deploy LivePerson's Conversational Cloud platform, shifting customer service from voice to WhatsApp and other digital channels. The rollout included a virtual assistant available 24/7, bot parking for inactive conversations, AI-generated conversation summaries for agents, secure-form-based identity verification, and LivePerson's Intent Manager to analyze customer intent and improve service-bot menu performance. The migration from Getnet's previous platform took just 20 days. Within 11 months, digital channels grew to 74% of support conversations (up from a smaller share, with voice dropping to 26%), CSAT rose from 45.91 to the mid-80s (an increase of over 85%), and NPS rose by 78 points on average across support menus that had previously been negative. As of July 2024, 17 bot automations supported 84 service groups and 300 human agents.
ID Logistics' SmartVision delivers 80% error reduction in warehouse order picking
ID Logistics launched SmartVision, an AI solution using stereovision cameras to interpret real-time data and monitor parcel preparation by warehouse operators. The system reduces picking error rates by 80% by continuously recording actions on the trolley and validating the number of parcels deposited in real time; recorded images also serve as verification for complaints. Team leaders get real-time visibility into quality rate, remaining volume, and palletizing conformity. The solution was first deployed at a food distribution site north of Paris, with plans to extend to retail sites in France and internationally.
Levi Strauss & Co. builds a Super Agent with Microsoft Foundry
Levi Strauss & Co. uses Microsoft Foundry and GitHub Copilot to build purpose-driven AI agents that connect to enterprise systems such as SAP, HR platforms and retail systems, orchestrated through a planned unified 'Super Agent' experience inside Microsoft Teams. The company first rolled out Microsoft 365 Copilot enterprise-wide, then used Foundry as an orchestration layer with Azure Functions for backend execution, aiming to let employees get answers and complete tasks like checking inventory or initiating HR requests through a single natural-language interface.
Val Verde Unified School District uses Microsoft Copilot to speed up IT and business operations
Val Verde Unified School District, a Southern California district of 19,000 students, piloted Microsoft Copilot and ChatGPT across administrative and business departments after its Board of Education directed staff to test generative AI. Director of Information and Instructional Technology Matt Penner used Copilot as a code assistant to fine-tune a custom state-reporting application that integrates data from the student information system, finishing the task in five hours instead of ten. About 150 employees were testing the AI tools, including HR staff improving job interview questions and staff summarizing hundreds of unopened emails into a manageable list.
Ørsted uses drones and Cognite Data Fusion for data-driven turbine maintenance
Offshore wind operator Ørsted uses drones to fly inside monopile wind turbines and collect images and sensor data, which is organized in Cognite Data Fusion, powered by Microsoft Azure, to run computer vision models that track corrosion and cracks. The data-driven approach lets Ørsted reduce the number of technicians needed for inspection from four to two and eliminates the need to physically enter confined turbine spaces, improving safety while reducing maintenance cost and downtime.
Franklin Templeton scales investment insight with Databricks Agent Bricks
Franklin Templeton built the internal SIGNALS application on Databricks Agent Bricks, combining proprietary fund scoring models with unstructured documents like prospectuses and fact sheets in Unity Catalog to generate AI-powered portfolio commentary. Analysts save 2+ hours per week, field teams surfaced $15 million in new product opportunities during early rollout, and AI-powered commentary coverage expanded from 200 products to hundreds.
Unilever scales AI across a 100% cloud-based operating model
Unilever's Chief Enterprise Technology Officer Steve McCrystal describes how the company has implemented more than 500 AI-based capabilities across the globe over the past decade, run on a fully cloud-based digital infrastructure that handles 8 petabytes of data across 25,000 daily data pipelines and 240TB of network traffic a week for over 3 billion transactions. Unilever has trained 23,000 employees in AI usage, opened an AI research hub in Toronto, and uses AI to help connect consumer demand, product development, planning, suppliers, logistics and manufacturing, achieving cost savings in its supply chain, governed by its Responsible AI framework.
Israel's Meteorological Service Predicts the Weather Using NVIDIA Earth-2
The Israel Meteorological Service integrated an AI weather model called F3, powered by NVIDIA Earth-2 and NVIDIA PhysicsNeMo (including the CorrDiff neural network), into its forecasting system to achieve hyperlocal 2.5-kilometer resolution precipitation forecasts. F3 achieved 90 percent lower computational costs compared to running a classic non-AI numerical weather prediction model on a CPU cluster, delivering forecasts in minutes rather than hours, and saved 20-30 percent of the Tel Aviv-Yafo Municipality's preparedness budget for things like urban flood safety. CorrDiff was trained to emulate IMS's regional high-resolution ICON model using low-resolution ECMWF IFS input, enabling four high-resolution forecasts per day. Partners including the Israeli Police, the National Fire and Rescue Authority, the Ministry of Transportation and the National Road Company have adopted the forecasts, and IMS used F3 to issue a high-resolution flood map during a January storm that dropped 150 millimeters of rain in six hours.
HP uses Databricks AI/BI Genie to enable self-serve analytics across the enterprise
HP's Analytics and Insights, product management and Digital GTM teams adopted Databricks AI/BI Genie to let business users ask natural-language questions of billions of telemetry events instead of waiting on engineers to build dashboards or write SQL, improving efficiency in analyzing product, sales and customer patterns by 40-50% and empowering thousands of users with self-service analytics, governed by Unity Catalog role-based access control.
From high costs to near real-time insights with Zerobus Ingest
Beusa Energy, a 30-year energy sector company spanning electric hydraulic fracturing, mobile power generation, electrical distribution, field gas processing and industrial manufacturing, migrated its high-frequency telemetry ingestion from a custom SQL Statement API pipeline to Databricks' Zerobus Ingest direct-write API. The migration required only swapping the write path (same .NET worker, same MQTT subscriptions), cutting ingestion cost from roughly 689 DBU per GB to about 0.29 DBU per GB, a 99% cost reduction, while eliminating the need for a separate streaming broker tier. Today more than 6,000 devices stream telemetry at 1 Hz across about 250 remote assets, ingesting ~22 million rows daily into a unified lakehouse with end-to-end latency of about three seconds, governed by Unity Catalog. Beusa is now using this high-frequency telemetry, combined with maintenance history, to train predictive-maintenance models forecasting remaining useful life of assets, moving from condition-based to predictive and eventually prescriptive maintenance.
Schaeffler unlocks factory insights with Microsoft Cloud for Manufacturing
Schaeffler, a global motion technology company with 10,000+ employees in Germany, worked with partner Avanade to pilot data solutions in Microsoft Fabric and a generative AI agent in Azure AI, part of Microsoft Cloud for Manufacturing, to unlock intelligence across diverse IT/OT systems (including ERP, SCM and MES) across a solution designed to scale to over 100 plants. Employees now ask the agent natural-language questions and get near real-time insights such as scrap rates, yields and energy usage; for example, asking what caused downtime on a line yields an instant detailed response instead of engineers spending hours investigating.
Amdocs Builds Generative AI Agents for Telecom
Amdocs built amAIz, a domain-specific generative AI platform helping telecom companies transform customer experiences, automate processes and optimize decision-making, using NVIDIA DGX Cloud, NVIDIA AI Enterprise software, NVIDIA NIM inference microservices, and NVIDIA Nemotron open-source reasoning models. amAIz agents, enhanced with NVIDIA's Llama Nemotron, autonomously handle complex multistep customer journeys spanning sales, billing and care. Using NVIDIA NIM microservices and telecom-based retrieval-augmented generation, Amdocs reduced tokens consumed by as much as 60 percent for data preprocessing and up to 40 percent for inferencing, and reduced query latency by approximately 80 percent.
Fifth Dimension unlocks insights and intelligence with Google Cloud
Fifth Dimension, founded in 2023, provides an AI platform combining machine learning, predictive analytics and natural language processing to uncover hidden patterns in real estate data, processing over 1TB of data per month by 2025. Facing scalability and cost problems with its original infrastructure, the company adopted Google Cloud, building its ML stack on Vertex AI (running Google Cloud Gemini and Anthropic Claude models), plus Cloud SQL, Cloud Run and Pub/Sub. This scaled processing capacity 50x to handle document-processing surges, supported 6x global client growth across multiple regions, and decreased infrastructure costs by 30% through serverless architecture. Model deployment time dropped from weeks to days, platform engagement tripled in 2025, and annual recurring revenue grew 6x in the same year.
Accelerating Large Language Model Inference with NVIDIA in the Cloud
Perplexity built pplx-api, an API for developers to integrate open-source LLMs with fast inference, served on Amazon EC2 P4d instances powered by NVIDIA A100 Tensor Core GPUs and accelerated with NVIDIA TensorRT-LLM (with a planned move to Amazon P5 instances with NVIDIA H100 GPUs). pplx-api achieves up to 3.1X lower latency and up to 4.3X lower first-token latency versus other deployment platforms, and switching external inference-serving API references to pplx-api lowered costs 4X, saving $600,000 per year. Using NVIDIA H100 GPUs and FP8 precision on Amazon P5 instances cuts latency in half and boosts throughput by 200 percent versus A100 GPUs in the same configuration. Perplexity also uses AWS's Kubernetes integration to scale elastically beyond hundreds of GPUs.
AES accelerates renewable energy adoption with Claude on Google Cloud
AES, a global energy company, built a multi-agent AI system using Claude on Google Cloud to automate internal safety audits of its renewable energy assets. The system uses document processing agents, task breakdown agents, and report generation agents to process hundreds of pages of safety documentation, evaluate compliance, and produce audit reports. Audit reports that previously took up to two weeks to complete (across roughly 1,550 internal audits annually) can now be generated in about an hour, a 99% reduction in time, alongside a 10-20% improvement in audit accuracy and a 99% reduction in audit costs.
Lippert Improves Customer Support with GenAI
Lippert, a $3.8 billion global manufacturer of components for RV, marine and automotive brands, used Databricks and Agent Bricks to consolidate fragmented data into a single lakehouse and deploy AI agents for customer support. An AI assistant trained on product manuals, technical case history and field-expert videos surfaces troubleshooting guidance for call center agents in real time, cutting new-agent ramp time from six months to four weeks (an 85% reduction). Using Agent Bricks' evaluation framework and synthetic data, model accuracy for the support agent improved from 33% to 84% within weeks. AI also analyzes thousands of support calls daily for coaching and quality scoring (versus a prior manual sample of 100 calls/month) and automates call summarization into Salesforce. The initiative is expected to save millions per year and reclaim hundreds of thousands of hours, with new agents planned for HR, warranty and supply chain.
HARTING reduces design time from weeks to minutes with Siemens and Azure AI
HARTING, a German maker of industrial connectors with more than 27,000 product options, built an AI-guided Han® Configurator powered by Microsoft Azure OpenAI Service and Microsoft Cloud for Manufacturing that lets customers describe requirements in natural language and get guided to the right product in about one minute, down from 15-20 minutes. For custom designs, a companion tool built by HARTING's Innovation Hub interoperates with Siemens' NX and Simcenter (part of Siemens Xcelerator) to refine 2D layouts, generate 3D models and run basic thermal FEM simulations, cutting custom prototype creation from weeks to minutes.
F.02 Contributed to the Production of 30,000 Cars at BMW
Figure deployed its Figure 02 humanoid robot at BMW Group Plant Spartanburg for an 11-month trial performing sheet-metal loading on an active assembly line, running 10-hour shifts Monday-Friday. Over the deployment, the robot loaded 90,000+ parts, logged 1,250+ hours of runtime, and contributed to the production of 30,000+ X3 vehicles, with a target cycle time of 84 seconds and placement accuracy goal of 99% per shift.
Albert Heijn revolutionizes the store employee experience using Azure OpenAI
Albert Heijn, one of the largest supermarket chains in the Netherlands with over 700 owned stores and more than 80,000 store employees, built a conversational assistant within its @AH Employee App using Azure OpenAI, hosted on Azure Kubernetes Service with data stored in Azure Database for PostgreSQL. Developed with Microsoft and partner EPAM, the assistant answers store employees' questions about restocking and customer requests, including a barcode scanner and store floorplan. It has been piloted in eight stores with plans to scale to 25 stores in summer 2025 and eventually all 700 owned stores.
Forcura Expedites Patient Care Using Generative AI on Amazon Bedrock
Forcura, a HITRUST-certified healthcare workflow management company serving over 900 healthcare providers representing about 1 million patients, added a generative-AI referral summary feature to its Referral Management product. Built on Amazon Bedrock using Anthropic's Claude 3 model family, the feature pulls key information such as patient demographics, clinical history and requested services into a concise summary. Forcura piloted the feature with three clients in April 2024 and reached general release within 90 days of the pilot start.
Workday Accelerates Generative AI / ML Product Development Using Amazon SageMaker
Workday uses Amazon SageMaker to let engineering teams build, train and deploy ML models, including LLMs, across AWS Regions to meet global customers' data-residency requirements. Using Amazon SageMaker Studio, Jumpstart and Ground Truth Plus, Workday piloted a closed-book ML application analyzing job descriptions, invoices and contracts, improving ML inference latency by a factor of five. Workday also received early access to Amazon Bedrock for generative AI prototyping. Head of Workday AI Shane Luke said the company has scaled from a thousand inference requests to tens of millions coming in daily.
Red Hat Helps DenizBank Transform Banking Services Through AI and Machine Learning Innovation
DenizBank, Turkey's fifth largest private bank with nearly 15,000 employees, adopted Red Hat OpenShift AI via its IT subsidiary Intertech to automate and standardize data science pipelines for its 120 data scientists working in risk management, marketing and customer relations. The platform reduced time-to-market for new microservices models from several days to just 10 minutes, enhanced credit qualification prediction models, and improved fraud prevention, while GPU usage was optimized through automatic scaling.
Harvard Medical School's Walter Lab runs 1.7 million protein interaction predictions using NVIDIA DGX Cloud
The Walter Lab at Harvard Medical School, with access to NVIDIA DGX Cloud through the National Science Foundation's NAIRR pilot program, completed nearly 1.7 million protein-protein interaction predictions on ColabFold in three months using 32-node DGX clusters with 256 A100 GPUs, a task that previously would have taken years. The team curated 40,000 high-confidence protein interactions among 300 human genome maintenance proteins using two custom ML tools (KIRC and SPOC), publishing results via the Predictomes website.
DXC Technology cuts platform TCO 30% and builds 11 AI agents using Databricks Data + AI Platform
DXC Technology, a global IT services company with fragmented systems from years of mergers, used the Databricks Data + AI Platform (Delta Lake, Unity Catalog, MLflow 3.0, and the Genie conversational assistant) to unify data across sales, HR, finance and project management. DXC built 11 AI agents (3 in production, 8 in pilot or development), reduced time-to-insight from months to days, and validated a 30% reduction in platform total cost of ownership.
Domyn builds Colosseum 355B, a sovereign AI foundation model, using NVIDIA DGX Cloud
Domyn (formerly iGenius), an Italian AI company serving highly regulated sectors such as financial services and public administration, used NVIDIA DGX Cloud with over 3,000 NVIDIA H100 GPUs to continue-pretrain Colosseum 355B, a 355-billion-parameter foundation LLM. Within one week Domyn had access to the dedicated infrastructure, and within two months completed continued pretraining, achieving 82.04% accuracy on the MMLU benchmark. The model powers Domyn's business intelligence agent, Crystal, a sovereign AI solution deployed on private infrastructure.
Nexo uses Agentforce to increase autonomous crypto customer support resolution by 70%
Nexo, a digital assets wealth platform serving over 7 million customers across more than 200 jurisdictions, built an Agentforce service agent to provide 24/7 FAQ support. Since launch, Agentforce has resolved 91,400 chats, a 70% increase in autonomous case resolution, saving Nexo's client care team 2,600 hours of manual work. Data 360 grounds responses in real-time data including LTV ratio tables pulled via a custom API, and Agentforce hands off to a human when a customer asks about a specific transaction to meet compliance requirements.
Rakuten accelerates development with Claude Code
Rakuten, a large Japanese ecommerce company, uses Claude Code to automate coding tasks across its engineering teams, reducing time to market for new features from 24 days to 5 days and achieving 7 hours of sustained autonomous coding on a complex open-source refactoring project. Rakuten also deployed Claude Managed Agents across product, sales, marketing and finance, plugging into Slack and Teams for non-engineers to complete tasks.
Delhivery achieves 160 ms latency for high-precision geocoding using Amazon EKS
Delhivery, a logistics provider in India, implemented a fine-tuned open-source Llama 3.2 1B large language model on Amazon EKS to support high-volume geocoding of pickup and drop-off addresses. The system processes up to 8,000 requests per minute at 160 milliseconds latency using NVIDIA A10G GPU-backed G5 Xlarge instances and the vLLM framework. Delhivery cut model-serving costs by approximately 80 percent and accelerated prototyping cycles from two days to under six hours, working with the AWS Prototyping and Cloud Engineering (PACE) team.
Robot Retasking in High-Mix Manufacturing with Workr
Workr, a manufacturing AI company, uses NVIDIA Omniverse, Isaac Sim and accelerated computing to let on-site operators retask industrial robots in under five minutes via a tablet interface, eliminating weeks of traditional programming. The edge AI pipeline runs on 2x NVIDIA RTX PRO 6000 Blackwell Max-Q GPUs attached to the robot cell, using models including RAFT-Stereo, Detectron2, NVIDIA Isaac FoundationPose and NVIDIA Isaac cuMotion, trained with synthetic data generated in Isaac Sim digital twins. Deployed with customers Yuasa International and Haas Alfex.
NVIDIA DRIVE AV Software Debuts in All-New Mercedes-Benz CLA
The all-new Mercedes-Benz CLA, the brand's first vehicle featuring the MB.OS platform, introduces advanced driver-assistance features powered by NVIDIA's full-stack DRIVE AV software, AI infrastructure and accelerated compute, enabling enhanced level 2 point-to-point driver assistance including urban navigation, active safety and automated parking. The CLA achieved a five-star EuroNCAP safety rating.
Austrian Academy of Sciences unlocks Ancient Greek with Mistral
The Austrian Academy of Sciences (OeAW), together with its Austrian Archaeological Institute, partnered with Mistral and services partner Reply to build Apollo, described as the first advanced large language model for Ancient Greek. Apollo is trained on a specialized corpus of 600 million words of historical Greek text plus tens of thousands of published inscriptions and papyri, helping researchers reconstruct damaged texts and identify thematic connections across collections. The OeAW reports Apollo turns work that once took years into hours, addressing over one million unread Greek papyri worldwide, with future phases planned for semantic search and handwritten inscription decipherment.
How Bouygues Construction Drives Sustainable Building Innovation With Digital Twins
Bouygues Construction, a global construction and infrastructure company based in France, built custom digital twin applications on NVIDIA Omniverse using the OpenUSD framework to unify 3D data from Revit, SolidWorks, and Cesium into a single collaborative environment. Using NVIDIA Isaac Sim, the company generates photorealistic synthetic training data for AI-based computer vision models covering rare or hazardous jobsite scenarios, and uses NVIDIA RTX PRO GPUs for real-time ray tracing. The approach delivered a 5x improvement in data integration speed and enabled safer job site planning through AI-driven construction site simulations and predictive scenario testing.
BNP Paribas: Transforming banking and empowering global employees with AI
BNP Paribas, the European Union's leading bank, has been leveraging Mistral AI's models for Global Markets use cases since Q3 2023, yielding positive results. BNP Paribas decided to expand its collaboration with Mistral to the whole group for 2024, with plans to explore Mistral's models on a broader range of use cases across its businesses.
France Travail's ChatFT AI Assistant Reaches Nearly 40,000 Agents as It Partners With Mistral AI
France Travail, France's public employment agency, has deployed several generative AI tools for its staff and job seekers. ChatFT, a generative AI assistant that reduces time agents spend drafting documents, has been in direct access for all France Travail employees since November 2024 and now has close to 40,000 agent users, more than 50% of France Travail's workforce. MatchFT helps advisers pre-select candidates through a guided conversation and contacts pre-selected candidates by SMS about a job offer. Neo, a module of ChatFT launched in October 2024, lets agents query France Travail's information system in natural language; it was tested in six agencies in Centre-Val de Loire plus roughly 120 volunteer users in La Réunion, and since April 15, 2025 has been progressively rolled out to four regions (Centre-Val de Loire, Pays de la Loire, La Réunion, Provence-Alpes-Côte d'Azur). On June 11, 2025 France Travail signed a partnership with Mistral AI: MatchFT and Neo already use Mistral models, with a goal of using Mistral 100% for these two tools by the end of 2025. France Travail also built a 'Calendrier de l'IA' AI-literacy program with Kokoroe and Microsoft, aiming to train 300,000 job seekers by the end of 2025.
CMA CGM Group deploys MAIA, an agentic AI platform co-developed with Mistral AI, across shipping, logistics and media operations
CMA CGM Group, in a five-year strategic partnership with Mistral AI signed in April 2025 that has embedded around twenty Mistral AI engineers in CMA CGM teams in Marseille and at CMA Media's headquarters, will progressively roll out 'MAIA, Powered by Mistral' starting June 1, 2026 -- an agentic AI platform designed to support nearly 80,000 employees across CMA CGM, CEVA Logistics and CMA Media. MAIA enables the development and orchestration of AI agents connected to business knowledge and internal applications, supporting content generation, document analysis, administrative task automation, and decision support. The Group has already deployed more than 55 AI projects and identified over 200 AI use cases, including AI-assisted agents that handle customer tracking and operational queries (escalating complex cases to humans) across more than 54 million customer requests a year, ETA prediction, smart booking and smart pricing, smart cargo release, vessel routing and energy-consumption optimization, demand and volume forecasting, and CEVA Logistics' 'smart labor' warehouse planning tools.
Deutsche Telekom Launches Industrial AI Cloud, One of Europe's Largest AI Factories, in Munich
Deutsche Telekom officially launched its Industrial AI Cloud in Munich's Tucherpark, a sovereign AI factory built over six months with NVIDIA and data center partner Polarise on nearly 10,000 NVIDIA Blackwell GPUs (including DGX B200 systems and RTX PRO Server GPUs), delivering up to 0.5 ExaFLOPS of computing power operated under German data protection, security and availability requirements. The factory is already operating at over a third of its capacity, with existing customers including Munich-based robotics company Agile Robots and engineering simulation firm PhysicsX. It underpins the 'Deutschland Stack' built with SAP (T-Systems providing infrastructure/platform, SAP providing Business Technology Platform and AI applications), and hosts the SOOFI (Sovereign Open-Source Foundation Models) research project, in which Leibniz Universität Hannover contracted Deutsche Telekom to provide infrastructure for training a sovereign, European-language-focused open-source LLM with around 100 billion parameters.
Digital Bank Debunks Financial Fraud With Generative AI
European neobank bunq, with more than 12 million customers and 8 billion euros of deposits, built an automated, AI-powered transaction-monitoring system to detect fraud and money laundering, replacing labor-intensive rules-based systems with supervised and unsupervised learning. Using NVIDIA GPUs, bunq accelerated its data processing pipeline more than 5x and, using the open-source NVIDIA RAPIDS suite of GPU-accelerated data science libraries, trained its fraud-detection model nearly 100x faster, improving model accuracy and reducing false positives. Bunq is also exploring NVIDIA NeMo Retriever, part of NVIDIA NIM inference microservices, to improve the accuracy of Finn, its personal AI assistant powered by a proprietary large language model.
TotalEnergies builds Pangea 5 supercomputer with Dell Technologies and NVIDIA
TotalEnergies is building Pangea 5, a next-generation supercomputer developed with Dell Technologies and NVIDIA that will increase the company's computing power to support seismic imaging, advanced simulation and AI-driven research in the energy sector.
City of Edinburgh Council cuts translation costs by 90% using Amazon Translate with AWS Partner City Trax
The City of Edinburgh Council, which provides services to over 500,000 residents including more than one in five who speak English as a second language, worked with AWS Partner City Trax to deploy an AI-powered document translation solution using Amazon Translate inside the council's secure AWS environment, with data residing in the UK to meet public sector data compliance standards. City Trax implemented the solution in 10 days, and within weeks 16 members of the customer services team were using it to translate documents into 75 languages within minutes. The council reduced translation costs from as much as £55 per document to £0.07, saving around £39,000 in fees within 10 months of going live, and cut translation delivery times from one week to 24 hours. Human translators still handle documents relating to legal processes as required by UK law.
BMW Group builds plant-scale digital twins with NVIDIA Omniverse, cutting vehicle aerodynamics simulation time 30x
BMW Group is building digital twins of its production facilities using NVIDIA Omniverse libraries, letting global production planning teams collaborate in real time, optimize the layout and design of complex manufacturing systems, and develop autonomous robot and vision AI applications prior to real-world deployment. Working with Siemens, BMW is also accelerating simulation of vehicle aerodynamics: tests on NVIDIA Grace Blackwell and CUDA-X-accelerated Simcenter Star-CCM+ software showed a 30x speedup for transient aerodynamics simulations of entire vehicle geometries, while reducing energy consumption and costs.
Kellogg Community College's admissions AI achieves 94% model-to-human agreement using Element451 Bolt Agents
Kellogg Community College used Element451's Bolt Agents to perform rubric-based first reads of admissions applications with human-readable rationales, achieving 94% model-to-human agreement and flagging 186 fraudulent applications in one week. The agent reads applications, verifies documents, scores based on institutional rubrics, and routes files for human review.
How Schibsted's AI model helped boost subscription sales
Schibsted developed a machine learning model to serve personalized subscription-sales recommendations to anonymous, non-logged-in front-page readers, using first-party demographic data (age/gender predictions from advertising) and sales insights to generate real-time, on-demand recommendations rather than batch processing. Of 158 candidate data points tested, around a dozen were retained as model features. A/B tests showed a 75 percent increase in subscription sales from front-page articles compared to previous models. The system is integrated with Schibsted's Curate content recommendation platform, using Flyte for orchestration and AWS DJL for model inferencing, and the team is transitioning to Tecton as a managed feature store.
Aramco's GenAI Foundation helps streamline upstream operations
Aramco built its GenAI Foundation, a strategic platform using generative AI across upstream technologies, on Red Hat OpenShift to centralize resources, streamline operations, and reduce costs at its Upstream Digital Center. The platform uses Red Hat OpenShift's containerized and microservices architecture and built-in observability tools to develop, deploy, and integrate large language models, and Red Hat Ansible Automation Platform to manage and configure virtual machines, reducing deployment times from weeks to hours. Drill and plant engineers complete tasks 10 times faster using AI to automate repetitive processes like data analysis, with each engineer and data scientist saving around 2 hours daily. The platform allows onboarding gen AI models in under an hour.
Virtual Human Meets Virtual Studio: AI Powers Text-To-Video for Broadcast
Hour One built the self-service platform Reals, which lets broadcasters automatically turn text into fully-produced videos with lifelike virtual presenters and virtual sets, accelerated by NVIDIA GPUs, CUDA and cuDNN. Customers include Berlitz, NBC Universal and DreamWorks. German sports site RAN used Hour One's 3D templates to create an immersive Sports News video series, and 24-hour broadcaster DeFiance Media built its own virtual news anchor, Raxana, using Hour One's platform.
Darrow raises $35M for an AI that parses public documents for class action lawsuit potential
Darrow, a legal tech startup founded in 2020, built an AI-based data engine that crawls thousands of sources of publicly available information (newsfeeds, social media, consumer complaints, regulatory filings, SEC filings, environmental reports, court dockets) to identify class-action litigation potential in areas such as data privacy violations, environmental contamination, and discrimination, and to predict likely legal outcomes and case value. A team of legal data specialists, including former lawyers, reviews the AI-surfaced insights before lawyers use the platform for case review and discovery. Around 50 law firms, covering hundreds of lawyers, were using the product, and active cases stemming from Darrow's data insights totaled around $10 billion in claims at the time of the funding round.
Ingka Group acquires Locus, an AI-powered logistics software company to transform the IKEA home delivery experience
Ingka Investments, the investment arm of Ingka Group (the largest IKEA retailer), acquired Locus, a USA-based logistics technology company offering an AI-powered logistics management platform with advanced route optimisation, real-time tracking, and smart use of vehicles and resources. The acquisition brings this technology in-house to enhance efficiency across IKEA's supply chain, from capacity management to last-mile execution, at a time when online sales reached 28% of total IKEA retail sales in FY24, up from 11% in FY19. Locus, described as an agentic Transportation Management System battle-tested in 350+ deployments across 30+ countries, will remain operationally independent and continue serving its own enterprise customers beyond Ingka Group.
How Serve Robotics Achieved 99.8% Success for Last-Mile Autonomous Delivery
Serve Robotics, which spun off from Uber in 2021, operates over 1,000 physical AI-powered sidewalk delivery robots serving over 2,500 restaurants across five cities. Its third-generation robots are simulated in NVIDIA Isaac Sim and powered by NVIDIA Jetson Orin edge AI platforms, achieving a 5x compute improvement over the prior Xavier generation and 12+ hours of battery life. The fleet has completed over 100,000 autonomous deliveries with a 99.8% completion rate, logging about 1 million miles of data monthly (nearly 170 billion image-LiDAR samples) to improve navigation and HD maps. The company plans to deploy 2,000 robots by the end of 2025.
KPMG: Improving project delivery speed with access to SAP software knowledge using SAP Joule for Consultants
KPMG member firms equipped around 7,000 SAP-specialized consultants with SAP Joule for Consultants, a generative AI solution delivered through the Joule copilot, giving consultants natural-language access to SAP's implementation knowledge base instead of manually searching documentation. KPMG also combined the solution with SAP Business Data Cloud to bring together structured and unstructured data from SAP and third-party systems, helping accelerate client research, summarize documentation, and streamline creation of transformation roadmaps and change-impact assessments.
AI-Media and ITV Announce Landmark Partnership to Revolutionize Captioning Services
ITV, the UK's largest commercial broadcaster, partnered with AI-Media to deploy its LEXI live automatic captioning solution across ITV's broadcast network, including its 18 regional news variants, using AI-Media's SDI Encoder Pro Cards to deliver captioning as VANC embedded data, plus AI-Media's Alta product for ITVX and other IP video workflows. The shift from a manual regional-news captioning workflow to LEXI's automated system is expected to significantly reduce staffing requirements and costs while enabling more captions, increased reliability and reduced latency; ITV's captioning team continues to develop bespoke topic models to maintain accuracy.
Vambe Powers Conversational Commerce Across Latin America with Claude
Vambe, a Chilean startup, built a conversational AI platform on Claude that automates sales conversations, lead qualification, and transactions for over 1,500 mid-sized businesses across Latin America via WhatsApp and Instagram. With Claude, Vambe achieved 95%+ multi-agent reliability (up from 30% with previous models), 40-60% higher conversion rates versus previous sales processes, 60% reduction in healthcare clinic no-shows, customer onboarding time reduced from 7 days to 3 days, and engineering debugging time dropped from 40% to 10% of workload, freeing 120 hours per month. AI-to-AI handoff reliability improved to over 95%, and average agents per platform increased from 1.0 to 2.3.
Lloyds Banking Group's Pioneering Use of AI at Scale Transforms Employee Experience and Operations
As part of a £4bn technology investment, Lloyds Banking Group rolled out Microsoft 365 Copilot starting with 300 licences via Microsoft's Early Access Program, scaling to a commitment of 40,000 licences by March 2025. A 1,000-strong network of volunteer 'flight-instructors' and weekly 'promptathons' supported adoption, reaching a 93% daily usage rate among licensed colleagues. Nearly 5,000 UK engineers use GitHub Copilot; one team converted 11,000 lines of code across 83 files, cutting a 65-hour task to 30 hours.
La Poste Empowers the Employee Experience with Microsoft 365 Copilot
La Poste, the French national postal operator with 233,000 employees, began deploying Microsoft 365 Copilot in 2023 to an initial group of 300 users, expanding to 2,000 across roughly fifteen job functions including HR, legal, marketing and communications. Employees use Copilot to summarise meetings, draft briefing notes, simplify legal documents, compare contracts, and generate social media content. The deployment was supported by partner Wavestone and a network of local ambassadors. 95% of employees want to keep their Copilot licenses.
ING Bank transforming operations through agentic AI
ING Bank is running multiple AI projects across operations, governed centrally under its COO. Live/early-production efforts include a retail chatbot, AI-driven identification of hidden affluent clients for marketing, AI-assisted transaction monitoring that helps investigators close standard alerts faster and focus on risk, and a customer due diligence (KYC) redesign that uses existing public and behavioural data to auto-answer most of a roughly 100-question review, cutting due diligence from days or weeks to seconds. ING is also developing agentic AI to handle mortgage applications end-to-end (credit checks, data collection) starting in 2026. ING's COO said AI introduced to an operations process yields a 25% productivity gain, with freed capacity redeployed to growth and more complex work, and its CTO described 'conservatively aggressive' governance restricting AI exploration to five areas under COO control.
Landmark government trial shows AI could save civil servants nearly two weeks each a year
Over 20,000 UK civil servants trialled Microsoft 365 Copilot for three months, using it to draft documents, summarise meetings, and handle admin tasks. The trial found it saved users an average of 26 minutes per day, equivalent to nearly two weeks per year. At Companies House, staff used Copilot to handle routine customer queries and speed up drafting responses and updating records. At the Department for Work and Pensions, Work Coaches used it to personalise advice for jobseekers.
Volkswagen Group Invests up to One Billion Euros in AI
Volkswagen Group announced plans to invest up to one billion euros by 2030 in AI-driven vehicle development, industrial applications and high-performance IT infrastructure, expecting efficiency gains and cost avoidance of up to four billion euros by 2035. More than 1,200 AI applications are already active across the Group. In vehicle development, Volkswagen is building an AI-powered engineering environment with partner Dassault Systèmes to shorten product development cycles to 36 months or less. On the shop floor, the Digital Production Platform connects more than 40 sites with AI applications for production optimization. The Group is also expanding its private cloud infrastructure to keep sensitive data processing within Europe.
How Hibbett Eliminated Out-Of-Stock Roadblocks With Similar Recommendations
Leading athletic-inspired fashion retailer Hibbett drove conversion by 2.4x by integrating Syte's out-of-stock suggestions, using Syte's visual-AI-powered product discovery solution to recommend similar items when a product is unavailable.
AT&T Drives AI Agents' Accuracy, Efficiency, and Performance With NVIDIA
AT&T built 'Ask AT&T' customer service AI agents and worked with implementation partner Quantiphi to use NVIDIA AI Enterprise, NVIDIA NeMo and NIM microservices to build a data-flywheel platform for continuous fine-tuning and evaluation. The pipeline uses NeMo Curator to clean training data, NeMo Customizer to fine-tune base models (Mistral 7B was selected as optimal), NeMo Evaluator to measure performance (Rouge, BERT F1), and NeMo Retriever for up-to-date retrieval, with models deployed as NIM microservices. AT&T reports up to 40% improvement in response accuracy after fine-tuning and an 84% decrease in call center analytics cost, and is collaborating with Arize AI to automate identification of difficult AI interactions.
Shell scales enterprise AI across 100+ applications with C3 AI
Shell has over 100 AI applications in various stages of development and deployment across its upstream, downstream, manufacturing and integrated gas businesses using the C3 AI Platform, including predictive maintenance scaled to 10,000 pieces of equipment. Shell's Real Time Production Optimization solution uses sensor data on pressure, temperature and flow rates to calculate optimum settings, squeezing out an additional 1-2% of LNG production and, at one LNG facility, cutting CO2 emissions by about 355 tonnes a day at full capacity (up to 130 kilotons per year). Shell renewed its strategic C3 AI partnership for five years in June 2021, and in 2021 Shell, C3 AI, Baker Hughes and Microsoft launched the Open AI Energy Initiative, an interoperable ecosystem of AI solutions for the energy industry.
Wyndham Hotels & Resorts builds Agentic Enterprise with Salesforce Data 360 and Agentforce
Wyndham Hotels & Resorts, which supports approximately 8,400 properties, used Salesforce Data 360, MuleSoft, Informatica, and Agentforce Service to create Wyndham Reservations Guest 360, a unified real-time guest record giving service reps access to reservation history, loyalty status, preferences, and sentiment plus AI-powered recommendations. Wyndham also launched agentic self-service support for franchise owners through the Wyndham Community online owner portal. Results included a 5% increase in Wyndham Rewards enrollment, 21% faster service conversations year-over-year, 20% autonomous resolution of franchisee inquiries, and a 25% reduction in contact center call volume.
Jabil pilots Azure Machine Learning and Project Brainwave for manufacturing visual inspection
Jabil, which provides advanced manufacturing solutions requiring visual inspection of components on production lines, piloted Microsoft Azure Machine Learning and Project Brainwave to improve inspection speed and accuracy, aiming to reduce workload and improve focus for human operators.
DHL Express deploys AI-powered DHLBot sorting robots in Singapore and South Korea
DHL Express introduced DHLBot, an AI-powered robotic arm developed in partnership with Dorabot that automates parcel sortation, installed in Singapore and South Korea. The DHLBot uses 3D and barcode cameras to scan airway bills and intelligently sorts packages into delivery bins by courier route. It sorts over 1,000 small parcels per hour with 99% accuracy, increasing overall operational efficiency by at least 40%.
Datasite's Agentforce resolves 70% of dealmaker chat inquiries autonomously
Datasite, an M&A SaaS platform used by over 1.8 million dealmakers across 180 countries, deployed Salesforce Agentforce to replace its Einstein chatbot for customer chat support. Agentforce pulls answers directly from Datasite's knowledge base of over 2,000 articles, transfers cases to live Assist team members with a chat summary when needed, and understands natural-language queries in multiple languages. Datasite expects Agentforce to resolve 70% of incoming chat queries, 20% more than the prior Einstein chatbot's 50%.
Lightwheel deploys NVIDIA GR00T N1.5 humanoid robots in Geely's live automotive factory using Isaac Sim simulation
Lightwheel built a simulation-first robotics platform on NVIDIA Isaac Sim, Isaac Lab, and NVIDIA Omniverse libraries to address data scarcity and the sim-to-real gap in physical AI development. For Geely, Lightwheel fine-tuned the NVIDIA Isaac GR00T N1.5 vision-language-action foundation model to the Unitree H1 humanoid robot's morphology using simulation-generated synthetic data and DexMimicGen data augmentation, then deployed it in Geely's live automotive factory where robots autonomously perform component transportation, precise part placement on inspection trays, and coordinated dual-arm manipulation for heavy components. The platform achieves a 100:1 simulated-to-real data ratio and reduced development cycles from months to weeks.
Safari365 boosts internal efficiency 30% with Salesforce Agentforce AI agents
Safari365, a bespoke African safari travel company, deployed Salesforce Agentforce AI agents to automate low-value, repetitive tasks such as reconfirming supplier bookings and generating personalized 'day notes' for customer itineraries. The company started with internal-use agents built with human-in-the-loop oversight, iterating using the Agentforce Testing Center. Impact included a 30% increase in internal efficiency, quotes delivered in 30 minutes versus 2-3 manually generated per day previously, and a six-week time-to-value from signing to live deployment.
Eggland's Best drives double-digit online unit sales growth with CommerceIQ AI-powered retail media management
Eggland's Best partnered with CommerceIQ to implement an AI-powered Retail Media Management strategy at one of its largest national retailers, refining advertising bid structures and keyword strategies guided by incremental ROAS (iROAS) metrics. The approach targeted consumers open to upgrading their egg choices and shifted retail media investment from generic keywords to detailed nutrient- and cooking-behavior terms. Results included up to a 33% increase in new customer engagement, high single-digit ROAS, and double-digit growth in online unit sales.
I-MED Radiology implements world-first artificial intelligence solution Annalise CXR
I-MED Radiology Network implemented Annalise CXR, an AI chest x-ray decision-support solution developed by annalise.ai (an I-MED joint venture with Harrison.ai), across its 250 clinics nationwide, used by 350 radiologists daily. Annalise CXR detects 124 clinical findings and was, at launch, proven to improve the accuracy of radiologists in identifying over 100 clinical findings, supporting over 38,000 patient cases per month.
GXO Pilots New AI-Driven Autonomous Vehicle in Warehouse Operations
GXO Logistics deployed its first autonomous industrial truck, powered by KION, at its facility in Épinoy, France, as a live pilot of AI-driven material handling. The pilot is part of a strategic collaboration between KION, NVIDIA and Accenture to bring AI innovation, digital twins and perception technology from simulation into real-world supply chain operations, aiming to show measurable improvements in cost-savings, productivity, operational scalability and safety.
IDEXX Laboratories cuts contract sanctions review from weeks to 20 minutes with Luminance AI
IDEXX Laboratories used the AI-powered contract automation platform Luminance to review its entire contract database for connections to sanctioned countries and entities following Russia's invasion of Ukraine, completing in 20 minutes a review that would typically have taken weeks manually. Separately, Koch Industries' in-house legal team uses Luminance Corporate to automate generation of NDAs, master service agreements and IT-related contracts, including self-service contract generation for non-legal business units, and Deloitte used Luminance on a project standardizing 4,500 contracts across 14 European entities for client BT in two weeks, estimating a 50% time savings versus manual review.
Volvo Autonomous Solutions and DSV launch autonomous freight operations in Texas
Volvo Autonomous Solutions (V.A.S.) and logistics provider DSV began commercial autonomous freight operations between Dallas and Houston, Texas, hauling the first commercial truckload with the Volvo VNL Autonomous truck integrated with the Aurora Driver self-driving system. The Autona/freight solution combines the purpose-built truck with self-driving technology from partners Aurora and Waabi plus fleet management and operations support; a safety driver remains onboard during this initial phase. V.A.S. has logged more than one million autonomous miles in regional and local freight since 2023.
Michelin runs 200+ AI use cases across manufacturing, supply chain and innovation
French tire manufacturer Michelin has more than 200 AI use cases in production, led by group chief data and AI officer Ambica Rajagopal. Its in-house IRIS system, protected by over 20 patents, partially automates end-of-line visual tire defect inspection to improve inspector efficiency and workplace ergonomics while operators retain final accountability. Machine learning forecasting tools improve demand forecast accuracy and proactively detect stock shortages in the supply chain. Michelin scans the startup ecosystem and uses tools including Databricks and Dataiku, and has partnerships with Microsoft and Rockwell Automation to codevelop AI solutions. The company reports AI-project ROI exceeding €50 million per year, growing 30-40% annually for three consecutive years, governed by an internal data office and responsible-AI principles (people-centric, explainable, accountable).
AXA Switzerland adopts Shift Claims agentic AI to transform claims handling
Shift Technology launched Shift Claims, an agentic AI solution that assesses and prioritizes insurance claims, guides claim handlers, and automates tasks across the claims lifecycle. AXA Switzerland is an early adopter, using it to support claims transformation and increase automation while keeping claims professionals informed for cases requiring human intervention. Shift reports early adopters of Shift Claims see 3% lower claims losses, 30% faster claims handling, a 60% overall automation rate, and greater than 99% accuracy in claims assessment.
Basque Government's EJIE builds high-speed AI translation tool on Red Hat OpenShift
EJIE, the Basque Government's IT management arm, built the Itzuli neural translator to translate between Spanish, French, English and Basque, replacing an older grammar-rule-based tool that struggled with long texts. Using Red Hat OpenShift to run GPU-backed containers for the neural translation models, with support from Red Hat Consulting and a Technical Account Manager, EJIE launched general and specialist legal/administrative versions of the tool. Usage grew 300% in three years: from about 2 million quarterly translation requests in early 2020 to over 6.7 million per quarter by August 2023, with peaks over 260,000 queries in a single day and over 30,000 daily users; official-service translation requests fell 40% over the same period as citizens used the self-service tool directly.
Stripe's ML flywheel cuts successful card-testing fraud attacks by 80%
Stripe built a machine learning-based system to detect and block card testing fraud, applying ML models at three levels of abstraction: overall prevalence estimation, identifying where attacks are occurring, and scoring individual transactions. A rapid data-labeling, retraining and redeployment pipeline, built on Stripe's Shepherd feature-engineering platform (developed with Airbnb) and the Flyte ML orchestration platform, lets the team react to new attack patterns within hours. The system is augmented by a large transformer model trained on billions of global transactions that generates embeddings used across card-testing detection use cases. Successful card-testing attacks on Stripe declined by 80% over two years even as Stripe's payment volume grew past $1 trillion.
PMI shifts 88% of phone and email conversations to messaging with LivePerson AI
Project Management Institute (PMI) moved its customer support model from phone and email to digital messaging channels using LivePerson's Conversational Cloud platform, reaching 88% of conversations on messaging channels. PMI's virtual assistant handles approximately 38% of incoming chats and logged 1.5 million interactions, while LivePerson's Copilot auto-suggests agent responses and summarizes conversations. Web messaging conversations grew 24% year-over-year to about 568,000 in 2023, and conversational AI's share of support volume rose from 4% in 2021 to 28.5% in 2024.
Marsh McLennan saves over 1 million hours with agentic AI assistant LenAI
Professional services firm Marsh McLennan built LenAI, an AI assistant now used by its roughly 90,000 employees for around 20 million requests annually. To fix intent-recognition problems in the original GPT-3.5-based assistant, Marsh McLennan worked with Predibase to fine-tune a Llama-3.1-8B-Instruct model using LoRA and Turbo LoRA, improving accuracy by 10-12% over GPT-3.5 and 7% over GPT-4o-mini, and cutting response time below 4 seconds. Since its December 2023 launch, LenAI has handled over 25 million cumulative queries and saved more than 1 million hours of team time in its first year.
UVeye uses vision language models to detect vehicle defects at scale
Automated vehicle-inspection company UVeye processes over 700 million high-resolution images each month and applies vision language models to convert this visual data into structured condition reports, detecting subtle defects, modifications or foreign objects. UVeye detects 96% of defects compared with 24% using manual inspection methods.
Open Universities Australia triples lead qualification with generative AI agents
Open Universities Australia (OUA) used LivePerson's AI Studio to build an LLM-powered generative AI agent that handles initial prospective-student inquiries, replicating the conversational style of human advisors. The agent achieved 3x lead qualification rates compared to self-searching students and 2x compared to OUA's prior scripted chatbot, with an average response time of 6.3 seconds versus roughly 2 minutes for human advisors.
DHL and Locus Robotics reach one billion warehouse picks milestone
Warehouse robots deployed by DHL Supply Chain and Locus Robotics completed their one billionth warehouse pick. Thousands of autonomous mobile robots operate in more than 40 DHL-managed facilities worldwide, coordinated through Locus Robotics' warehouse execution platform. DHL reports 30-180% increases in units picked per hour and an 80% reduction in training time for warehouse staff, and has an expanded agreement to deploy 5,000 AMRs globally.
NPG launches AI-powered 'Frag Mich' article chat for subscribers
Neue Pressegesellschaft (NPG), publisher of Südwest Presse, Märkische Oderzeitung and Lausitzer Rundschau, launched 'Frag Mich', a question-answering chat feature built with AI vendor Retresco. Subscribers can ask free-form questions about current and archived articles; a RAG system answers using only editorially reviewed content and cites sources.
Fastweb + Vodafone: Transforming Customer Experience with AI Agents using LangGraph and LangSmith
Fastweb + Vodafone, part of the Swisscom Group and one of Europe's leading telecommunications providers serving millions of customers across Italy, built two AI agent systems on LangChain and LangGraph. Super TOBi transforms its existing chatbot TOBi into an agentic system using a Supervisor agent (for routing and guardrails) and specialized Use Case agents following the LLM Compiler pattern to call customer APIs and execute transactional actions like offer activation or payment updates. Currently serving nearly 9.5 million customers through the Customer Companion App and voice channels, Super TOBi achieves a 90% correctness rate, an 82% resolution rate, and a Customer Effort Score of 5.2 out of 7. Super Agent is an internal tool that augments call center consultants with diagnostics and guidance, storing operational knowledge as a graph in Neo4j and using an automated ETL pipeline with LangGraph and LLM agents to convert business-authored procedures into the knowledge graph; it has driven One-Call Resolution rates above 86%. The company uses LangSmith for observability, daily automated evaluation of chatbot responses, and continuous improvement.
United Imaging Healthcare Enhanced the Speed and Quality of MR Imaging
United Imaging Healthcare worked with NVIDIA since 2018 to accelerate its magnetic resonance (MR) imaging products with AI, developing an AI-assisted compressed sensing (ACS) technology to speed up MR image acquisition and reconstruction. Using NVIDIA data center GPUs, the CUDA Toolkit, cuSolver matrix computation library, and NVIDIA AI Enterprise software (with support from NVIDIA DevTech engineers and solution architects), United Imaging Healthcare reduced image reconstruction time by nearly 95%, with NVIDIA's support teams further accelerating algorithms by an additional 3x, on top of an initial 10x increase in computational speed from a tailored algorithm.
Delivering more for our riders in a year of incredible growth
Waymo's autonomous ride-hailing service scaled significantly in 2025, beginning to serve over 1 million fully autonomous rides every month starting that spring, and serving over 14 million trips in 2025 alone (more than tripling public rides from the prior year), putting it on track to exceed 20 million lifetime trips by year end. Riders spent over 3.8 million hours in Waymo vehicles. The all-electric fleet helped riders avoid over 18 million kilograms of CO2 emissions, 3x more than the prior year, while achieving more than a 10-fold reduction in serious injury or worse crashes compared to human drivers. Waymo plans to expand ride-hailing operations to over 20 additional cities in 2026, including international cities like Tokyo and London, and is expanding freeway riding and airport service (San Jose Mineta, San Francisco International, and Miami International airports).
Coca-Cola brings its iconic Santa into the future with an AI-powered avatar
Coca-Cola created an AI-powered avatar of its iconic Santa character that talks with consumers in real time, using Microsoft Foundry, Azure Speech in Foundry Tools, and Azure Cosmos DB. The avatar launched in 26 languages and spread across 43 countries, generating over 1 million consumer engagements.
Tawuniya transforms CX with unified customer intelligence
Tawuniya Insurance, the largest insurer in the MENA region by gross written premium, implemented Sprinklr Service and Sprinklr Insights to unify customer conversations, sentiment, and service operations from social media, messaging, and other digital channels onto a single platform. The company used sentiment analysis to turn customer feedback into actionable insights and automated key elements of the customer service journey including routing, queue management, and customer communications. Results: 158% improvement in First Contact Resolution (from 31% to 80% within 90 days), a 32-point increase in CSAT (from 51 to 83), and average response times reduced from 45 minutes to 56 seconds.
Foxconn Develops Physical AI-Enabled Smart Factories With Digital Twins
Foxconn (Hon Hai Technology Group) uses physically accurate digital twins integrating NVIDIA Omniverse libraries and OpenUSD to design, deploy, and manage high-volume production facilities, including those producing NVIDIA GB200 Grace Blackwell Superchip systems. Its Fii Omniverse Digital Twin (FODT) platform creates virtual replicas of factories, enabling simulation-driven design, real-time monitoring, and optimized operations. Using NVIDIA PhysicsNeMo AI models, Foxconn achieves 150x faster computational fluid dynamics simulations for thermal analysis (minutes vs. hours). Standardized OpenUSD-based digital twin assets enable rapid migration of entire production lines between global factories (e.g., Taiwan to Mexico). Robot workcells and AGV logistics are simulated in FODT before physical deployment: complex robotic tasks such as screw tightening and cable insertion are simulated and refined with NVIDIA Isaac Sim, Isaac Lab, FoundationPose models, and NVIDIA cuMotion, while AGV path is optimized by connecting simulations with Material Control Systems and NVIDIA cuOpt. Foxconn also built video analytics AI agents with NVIDIA Metropolis and the NVIDIA AI Blueprint for video search and summarization to monitor factory floors, and developed FoxBrain, an AI platform powered by NVIDIA NeMo trained in four weeks, described as Taiwan's first large language model with advanced reasoning capabilities. Leo Guo, General Manager of Fii Robotic Group, said the company believes it can cut factory setup and planning time by about 50%.
Driving manufacturing excellence at dsm-firmenich with AI and standardized PI Server data
dsm-firmenich implemented AVEVA PI Server and Asset Framework globally to establish an Operations Data Layer (ODL), a standardized data platform supporting 3,500 users, using the OEE (Overall Equipment Effectiveness) KPI to drive business value. Process data from the PI Server is contextualized in Shiftconnector plant process management to help manufacturing teams drive continuous improvement in batch and continuous operations, from active ingredients manufacturing to filling and packing lines. Collected and contextualized performance data preserves operational knowledge, which can then be accessed through an AI copilot supporting industrial workers in troubleshooting, root cause analysis, and onboarding new workforce.
How Schiphol is leveraging tech, design, data and AI-powered intelligence to redefine airport capacity and flow management
Royal Schiphol Group built in-house AI-powered tools — Dynamic Time Slots, (Passenger) Flow Balancing, and Gate Planning Insights — to align passenger demand with security and gate capacity. Dynamic Time Slots lets passengers pre-book security check times, steering an average of 7.5% of passengers away from peak moments (up to 22% on critical days), with 99.6% of passengers waiting under 10 minutes. Flow Balancing uses real-time sensor data, predictive models and simulation (a digital twin) to steer passenger groups, preventing over 4,000 minutes of crowding in 2024. Gate Planning Insights uses machine learning to optimise gate stand planning and turnaround operations. Overall passenger satisfaction score is 4.4 out of 5.
Pony.ai Realizes Gen-7 Robotaxi City-wide Unit Economics Breakeven
Pony.ai's Gen-7 Robotaxi reached city-wide unit-economics breakeven in Guangzhou in November 2025, with daily average orders per vehicle reaching 23, following official launch of fully driverless commercial Robotaxi operations in Guangzhou, Shenzhen and Beijing. The company operated 961 Robotaxi fleet vehicles (667 Gen-7 units) as of November 23, 2025, targeting 3,000+ vehicles by end of 2026, and reported an additional 20% reduction in Gen-7 autonomous driving kit bill-of-materials costs for 2026 production versus the 2025 baseline. Robotaxi services revenue grew 89.5% year-over-year in Q3 2025, with fare-charging revenue up over 200%.
Hyundai Capital America: Transforming customer experience with messaging innovation
Hyundai Capital America (HCA), the finance arm of Hyundai Motor Group for Hyundai, Kia and Genesis leases and loans in the US, partnered with LivePerson starting from a 2018 proof of concept to shift from a voice-dominated contact center to a data-driven, messaging-first strategy across SMS and Apple Messages for Business. Using deflection-to-messaging on high-friction payment interactions, HCA achieved 94% customer satisfaction, an 89% containment rate, and grew messaging engagement from 10,000 to over 100,000 customers served monthly. HCA and LivePerson co-developed a PCI-compliant chatbot debit card payment integration with payment processor ACI.
C.H. Robinson Scales Fleet of AI Agents Past 30
C.H. Robinson has deployed more than 30 proprietary AI agents at scale across its global logistics network, spanning the shipment lifecycle from quoting to tracking. The agents have automated over 3 million shipping tasks and contributed to a 30% increase in productivity since 2023. Named agents include a Quoting Agent (32-second price quotes, 1M+ quotes), an Orders Agent (processes 5,500 truckload orders/day, saving 600 labor hours daily), an LTL Classifier Agent (75%+ of LTL orders automated, up from 50%), an Appointments Agent (3,000+ appointments/day across 43,000 locations), a Truck Posting Agent, and a Shipment Tracking Agent.
How Walmart Achieved Enterprise Transformation Through Digital Experience Innovation
Working with implementation partner Nativa on LivePerson's technology, Walmart customized and integrated conversational technology into its WhatsApp channel, including a FAQ-Transactional Bot to streamline searches, inquiries and purchases. The solution delivered a 60% increase in productivity and a 10 percentage point improvement in CSAT/NPS, along with operational savings through automated customer service processes.
Capitec Bank scales conversational banking on WhatsApp with LivePerson
Capitec Bank, one of South Africa's largest retail banks, worked with LivePerson to bring banking services onto messaging platforms like WhatsApp, so clients did not need to delete banking apps to save phone storage. Capitec's client base grew from 16 million to 22 million, it handled 15 million client conversations in 2023 (7 million via WhatsApp), and it deployed LivePerson's Generative Insights and Agent Assist tools to help staff manage a larger volume of interactions, reporting 78% CSAT and 71% first contact resolution.
How Salesforce Deployed Agentforce to Answer Questions and Capture Leads on Their Website
Salesforce deployed Agentforce, agentic AI powered by Data 360, on salesforce.com to answer visitor questions instantly, grounded by more than 2,000 web pages and 400 product records, and to capture leads for Sales Cloud when visitors want to talk to sales. Since launching in February 2025, Agentforce has handled over 100,000 conversations, supports 11 languages, generated more than 30,000 new leads, and reduced time to qualify opportunities by 40% year over year.
Investec bankers save 200 hours a year with Microsoft Copilot for Sales
Investec, an international bank and wealth manager, was the first Microsoft customer to go live in production with Microsoft Copilot for Sales, rolling it out to 900 bankers in the UK. Copilot generates client history summaries and AI-powered email drafts inside Outlook and Teams, saving roughly 200 hours a year across the bank and shaving about 10 seconds off each post-call CRM update. Investec is extending the rollout to 700 bankers in South Africa.
Bayer uses generative AI and machine learning to speed crop breeding and agronomy
Bayer uses AI and data science across its agriculture business. Its precision breeding platform, combining machine learning and digital twin technology, has cut a breeding cycle from five to six years down to four months. Bayer trained a large language model called E.L.Y. on internal agronomy data to answer farm management questions, and uses AI models to guide planting density and irrigation decisions on its own seed-production fields, increasing field productivity by over 30%; it also offers AI-powered seed scripts and planting-density recommendations to growers through its FieldView digital farming platform, which manages more than 220 million acres in over 20 countries.
Hyundai Motor Group builds NVIDIA Blackwell AI factory for mobility and smart factories
Hyundai Motor Group is building an AI factory using 50,000 NVIDIA Blackwell GPUs to train, validate and deploy AI for in-vehicle systems, autonomous driving, smart factories and robotics. The collaboration includes an approximately $3 billion investment with the Korean government to build a national physical AI cluster, an NVIDIA AI Technology Center and Hyundai's Physical AI Application Center. Hyundai is exploring using NVIDIA Omniverse and Cosmos on RTX PRO Servers for factory digital twins, and using NVIDIA DRIVE AGX Thor running DriveOS for driver-assistance and in-vehicle AI.
Nexo resolves 62% of customer inquiries with Salesforce Agentforce
Digital assets wealth platform Nexo deployed Salesforce Agentforce across its app and website to handle more than 20,000 complex customer inquiries a month requiring real-time financial data and jurisdiction-specific knowledge. Agentforce now autonomously resolves 62% of cases, saving 2,600 hours of manual work, and hands off unanswerable questions to human reps. Nexo built separate global and UK sandboxes to test jurisdictional compliance and created a custom object to translate crypto-specific terms for regulatory accuracy.
Welt TV launches AI-generated news show with a synthetic avatar presenter
Welt TV, owned by Axel Springer, launched a five-minute AI-generated news programme hosted by a synthetic avatar modelled on reporter Paul Klinzing. Around 90% of the programme, including topic selection and scriptwriting, is produced using AI. Axel Springer has also rolled out WELTgo!, a GPT-based assistant for subscribers, and uses AI to repurpose older broadcasts for new platforms.
Phagos uses generative AI on AWS to match bacteriophages to bacterial infections
Phagos, a Paris-based biotech startup, uses generative AI models built with Amazon SageMaker AI to match bacteriophages to target bacteria for phage therapy, replacing a manual trial-and-error process. The AI models cut wet-lab testing needs by 50% and reduced phage-candidate screening time by 99.5%, from 29 hours to 10 minutes per bacteria. Phagos has treated more than half a million animals in France and can now develop a new treatment in two months versus 10+ years for traditional antibiotic development.
Mayo Clinic deploys NVIDIA DGX SuperPOD to accelerate pathology foundation models
Mayo Clinic deployed an NVIDIA DGX SuperPOD with NVIDIA DGX B200 systems to support foundation model development for pathomics, drug discovery and precision medicine. In partnership with Aignostics, Mayo Clinic built the Atlas pathology foundation model, trained on more than 1.2 million histopathology whole-slide images. The new infrastructure is reducing four weeks of pathology slide analysis work to one week.
Spotlight: Qodo Innovates Efficient Code Search with NVIDIA DGX
Qodo, a multi-agent code integrity platform, built its AI agents on retrieval-augmented generation powered by a state-of-the-art code embedding model trained on NVIDIA DGX. Qodo fine-tuned two embedding models, Qodo-Embed-1-1.5B and Qodo-Embed-1-7B (based on Qwen), achieving state-of-the-art accuracy on the Hugging Face MTEB CoIR leaderboard in their size categories. In a collaboration with NVIDIA, Qodo's code indexer, RAG retriever, and embedding model were substituted into NVIDIA's internal RAG solution (Genie) for searching private code repositories, integrated into NVIDIA's internal Slack system, yielding more detailed and accurate responses to technical questions from expert C++ developers than the original pipeline, evaluated using Ragas-generated synthetic questions against RTXDI, RTXGI and RTXPT SDK repositories.
Najm's Customer Experience Transformation with WhatsApp
Najm, a motor vehicle insurer in Saudi Arabia, partnered with BAB, LivePerson's local Saudi partner, to transform customer support from voice-only to AI-powered chatbots on WhatsApp, including IVR deflection to WhatsApp. The transformation reduced response times from three hours to 15 minutes, cut contact center costs by 60%, and now handles 250,000 conversations per month, with 40% of conversations occurring on WhatsApp/messaging versus phone.
BMW Group deploys humanoid robots in production with Hexagon and Figure AI
BMW Group carried out an initial test deployment of Hexagon's humanoid robot AEON at its Leipzig, Germany plant in December 2025, part of a phased program with a further test deployment planned for April 2026 ahead of the actual pilot phase starting summer 2026. This follows an earlier, successfully completed pilot with Figure AI's Figure 02 robot at BMW's Spartanburg, USA plant: over ten months, Figure 02 worked ten-hour shifts five days a week, supporting production of more than 30,000 BMW X3 vehicles by removing and positioning sheet metal parts for welding, moving more than 90,000 components across roughly 1.2 million steps and 1,250 operating hours. BMW created a new 'Center of Competence for Physical AI in Production' to evaluate and scale humanoid robotics -- combining digital AI agents with physical robots -- across its production network.
PwC Saves USD$150M Modernizing Across 136 Countries with Microsoft 365 Copilot
PwC rolled out Microsoft 365 Copilot to its global workforce, providing secure, intuitive AI-powered tools. The firm saved more than USD$25 million by consolidating platforms and USD$150 million in valuable employee time through Copilot use, in a rollout spanning 136 countries.
Ping An's AI-Driven Car Insurance Claims Processor
Ping An launched a credit-based mobile app for car insurance claims, letting drivers report claims, upload photos for AI-assisted damage assessment, and receive compensation online. The system is built on an AI smart claims credit model using image damage assessment, OCR bill identification, and biometrics. Smart scheduling provides surveying within 5-10 minutes for 95.5% of claims, 30% of cases use AI image recognition for loss assessment with 95% accuracy, and compensation can be received within seconds via face recognition.
How 7-Eleven, Inc. Built a Game-Changing GenAI Creative Assistant for Marketing with Databricks
7-Eleven's AI Center of Excellence built a multi-agent GenAI marketing assistant using LangGraph and Databricks, moving beyond early document-based chatbots and basic RAG. The system includes a Campaign Creative Generator, Copywriter Bot, General Assistant, and Supervisor Agent that orchestrates workflow and reviews outputs, with human-in-the-loop review, web search for trend intelligence, and guardrails filtering toxicity, PII and off-policy responses. Databricks provides governance and observability via MLflow tracking, automated agent evaluation, and inference tables. The assistant reduced manual creative hours, with campaign concepting and scripting that used to take hours or days now generated, refined, and approved in minutes; user feedback included calling it 'a game changer' and 'better than ChatGPT.'
From lab to newsroom: How Reuters builds AI tools journalists actually use
Reuters developed a suite of AI-powered newsroom tools with a human-in-the-loop approach: Fact Genie, an AI-assisted summarisation tool that scans entire documents in under five seconds to suggest newsworthy alerts (helping Speed teams publish the first alert within about six seconds of a press release, across roughly 100,000 business news alerts published monthly by 250-300 journalists); LEON, an AI-powered headline assistant; and AVISTA, a machine-learning tool for sourcing, tagging and archiving photos and videos. Fact Genie took about four months from prototype to production rollout, starting with senior journalists in the US and UK before expanding, and the team also uses AI to filter out non-newsworthy content before it reaches a large language model, cutting alert-generation time from about a minute with GPT-3.5 Turbo to about 10 seconds with GPT-4o mini.
PostNL embeds AI-driven route planning and sales intelligence across its logistics platform
PostNL's 2025 transformation embedded artificial intelligence and advanced analytics throughout operations: predictive algorithms support real-time route planning, parcel flow optimisation and capacity forecasting, and a new AI-powered sales intelligence tool converts ideal customer profiles into prioritised prospect lists with detailed shipment insights for account managers. In Belgium, PostNL piloted a fully digital onboarding journey for small business customers, reducing registration time from 15-20 days to a maximum of one day.
Adyen launches Uplift, an AI payment optimization suite piloted by 60 enterprise businesses
Adyen Uplift is an AI-powered payment optimization suite trained on Adyen's global transaction dataset of over $1 trillion in payments data, helping businesses increase payment conversion, simplify fraud management and reduce payment costs. The pilot, run with businesses including Patagonia, On, Indeed, NordSecurity and Fubo, showed up to a 6% uplift in payment conversion, up to 5% lower payment processing costs, and an 86% average reduction in manual fraud rules for pilot customers.
Balfour Beatty invests 7.2 million pounds in Microsoft 365 Copilot and AI 'smart agents' for construction quality checks
Balfour Beatty announced a 7.2 million pound investment in Microsoft 365 Copilot, one of the largest AI investments of its kind in UK construction and infrastructure. Balfour Beatty is also partnering with Microsoft to develop AI-powered 'smart agents' to improve Quality, Health & Safety and Assurance processes; the first trial is on the A9 Dualling: Tomatin to Moy project in Scotland, automating early-stage reviews of Inspection and Test Plans that traditionally took several hours.
BMW Group scales Virtual Factory digital twins to cut production planning costs by up to 30 percent
BMW Group is industrialising its Virtual Factory, with production planners continuously scaling applications in digital twins of over 30 production sites built on NVIDIA Omniverse. Automated collision checks for new vehicle models that once took almost four weeks of real-world testing now take three days to simulate virtually, and production planning costs are projected to fall by up to 30 percent.
General Motors uses AI for digital twins, battery leak detection, and weld/paint inspection in manufacturing
GM builds in-house AI solutions used across its manufacturing facilities: digital twins that simulate a running production line before it's constructed, AI/machine learning that quickly and accurately pinpoints potential leaks in battery packs, and a combination of robotics and a proprietary AI tool that inspects welds and paint coats to identify irregularities and anomalies.
Miami Dade College boosts student pass rates by 15% with Microsoft 365 Copilot
Miami Dade College deployed Microsoft 365 Copilot among its President's Cabinet and senior leaders, and used Microsoft Copilot Studio to build AI-powered assistants supporting personalized learning, coding exercises, essay structures and math problems around the clock. In the first semester of implementation in STEM and advanced analytics courses, the college saw a 15% increase in student pass rates and a 12% decrease in dropout rates. A survey found 77% of users completed tasks faster, 76% saw improved work quality, and 81% reported increased productivity; if just 15% of MDC's more than 6,000 employees saved 12 minutes a day, that would total nearly 50,000 hours a year. Encouraged by these results, MDC expanded its Copilot deployment by 400 licenses, prioritizing student-facing employees.
Far EasTone Telecom (FET) is turning agentic AI into real operations impact
Far EasTone Telecom (FET), a Taiwanese operator, applied Microsoft's network operations agent (NOA) reference architecture to redefine cloud-native network operations by embedding agentic AI across its network operations center (NOC) and change-management workflows. Nearly 60% of FET's NOC operations are now AI-assisted, with about 10,500 operational tasks executed per month, including incident summaries, automated ticket closure, network checks, and proactive voice notifications. AI agents handle large-scale alarm correlation and root cause analysis in seconds, supporting nearly 7,000 monthly operational queries with an average response time of 16 seconds, and most maintenance actions complete within one minute, significantly reducing human error and accelerating recovery times.
Volkswagen Autoeuropa's Welding Online Monitoring System predicts weld failures
At Volkswagen's Autoeuropa plant in Palmela, Portugal, engineers developed the Welding Online Monitoring System for about €17,000 ($17,850), using a custom algorithm to analyze real-time operational data from welding guns. The system predicts abnormal behaviour and potential failure before quality problems appear on the production line.
Southeastern Railway cuts case handling time 10.8% with ServiceNow Now Assist for CSM
Southeastern Railway, working with ServiceNow partner UP3, deployed ServiceNow's Now Assist for CSM to automatically summarize customer chats and cases for its Contact Centre Agents. Since 1 April 2025, Southeastern has routed 27,500 chats to live agents; average case handling time fell by 1 minute 30 seconds, a 10.8% reduction, saving approximately 687.5 hours of agent time (85+ full working days). Southeastern's existing ServiceNow Virtual Agent already handled around a third of customer queries, resolving 80% of those autonomously, with each Virtual Agent interaction 825% more cost-effective than a phone call.
Neue Pressegesellschaft launches AI article chat 'Frag Mich' with Retresco
German media group Neue Pressegesellschaft (NPG), whose news portals get around 24 million visits a month, launched 'Frag Mich' ('Ask Me'), an AI question-answering feature built with Retresco using Retrieval-Augmented Generation (RAG) technology. Available in the mobile apps of NPG titles SÜDWEST PRESSE, Märkische Oderzeitung and LAUSITZER RUNDSCHAU, it lets subscribers ask free-form questions and get answers grounded exclusively in editorially reviewed articles, with clearly cited sources and links to original content.
MITRE's Federal AI Sandbox accelerates government AI research with NVIDIA DGX SuperPOD
Nonprofit MITRE built the Federal AI Sandbox, powered by NVIDIA DGX SuperPOD, to give federal sponsors an affordable, centralized space to test and deploy AI and machine learning across domains such as weather forecasting, cybersecurity, and public benefits administration. The DGX SuperPOD delivers a 300-fold performance increase over MITRE's previous AI computing capabilities and supports thousands of researchers; MITRE is using it with NVIDIA Omniverse and NVIDIA Earth-2 to develop 1-kilometer precision weather forecasts with NOAA and the National Weather Service, plus a foundational model for cybersecurity threat analysis across 190 nations.
Wiley boosts service efficiency with Salesforce Agentforce
Publisher Wiley uses Salesforce Agentforce alongside Service Cloud and Einstein AI to manage customer service, letting AI agents resolve common issues like password resets and account access so human reps can focus on complex cases. Agentforce improved case resolution by over 40% compared to Wiley's previous chatbot, helped onboard seasonal agents 50% faster, and delivered a 213% return on investment with $230,000 in annual cost savings.
Nordea's Nova chatbot usage doubles in two years
Nordea's chatbot Nova (Nordea Virtual Agent), in service since 2017, is expected to handle close to seven million customer chat sessions in 2024, a 100% increase over two years. Nova answers simple banking questions in two seconds and uses 'deep links' that guide customers directly to a feature, such as viewing a card PIN. AI Trainers add at least 20 phrasings per intent, plus 10 tricky variations, to keep Nova understanding a wide range of customer questions.
KLM deploys AI to reduce in-flight food waste
KLM uses a machine-learning model called TRAYS, developed with Kickstart AI, to predict how many booked passengers will actually board a flight so its Meals On Board System can calculate the exact number of meals needed. The prediction runs from 17 days before departure until 20 minutes before departure and has cut food waste by up to 63% based on the expected number of passengers per flight, saving more than 100,000 kg of meals a year across flights catered from Schiphol.
Automation accelerates AstraZeneca's mission to save millions of lives by releasing 20 new medicines by 2030
AstraZeneca partnered with ServiceNow to replace manual research and administrative tasks with the ServiceNow AI Platform as part of its goal to launch 20 new medicines by 2030. A digitized, automated, AI-powered equipment and supplies ordering system on the AskAZ employee portal cut tasks that previously took up to 30 minutes, such as handwritten lab supply orders, down to seconds; more than 60,000 requests across 400+ categories now flow through the platform annually, saving an estimated 30,000+ hours per year. AstraZeneca also built "Onboarding 2.0" on the ServiceNow AI Platform, with personalized portals for new hires; automating roughly 10% of the 50+ hours managers previously spent onboarding each employee is projected to save over 90,000 hours as the company recruits more than 20,000 new employees.
Albert Heijn's AI-powered assistant Steijn helps millions of customers decide what's for dinner
Albert Heijn, part of Ahold Delhaize, built the AI-powered assistant Steijn, which helps millions of customers answer the daily question "What's for dinner?" by suggesting personalized recipes and step-by-step guidance. Ahold Delhaize is also scaling its proprietary retail media platform, Edge, which enables on-site display, sponsored search, and in-store digital screens, to its U.S. brands after delivering strong results across several European brands.
AI serves QoS & operational efficiency
Orange France deployed AI-based tools across its network operations to improve quality of service and operational efficiency, including the Recovisu diagnostic tool that helps technicians correct errors during interventions, automated frequency declarations with the National Frequency Agency, AI-driven preventive maintenance for equipment such as antenna batteries, and an AI-evaluated citizen infrastructure-damage-reporting app. The NRJ5G app analyzed consumption data to prevent unnecessary mobile site electrical subscription upgrades, saving nearly 48 million.
McKinsey's Firmwide Knowledge Copilot "Lilli"
McKinsey launched Lilli, a generative-AI platform that provides streamlined, impartial search and synthesis of the firm's internal knowledge to help consultants build client engagement plans. The firm's intellectual property spans more than 40 curated knowledge sources today, and McKinsey said there will be more than 100,000 documents and interview transcripts of internal and third-party content in the future. A user asks a question and Lilli scans the firm's knowledge base, surfaces five to seven of the most relevant pieces of content, summarizes key points, and identifies subject-matter experts. McKinsey piloted Lilli with real users and was scaling it from an initial team of three to more than 70 experts, aiming to reach thousands of colleagues by the end of the year.
Admiral Seguros: Tractable delivers outstanding customer services through touchless claims
Admiral Seguros, a Spain-based auto insurer serving over 260,000 clients, integrated Tractable's visual AI to enable Straight-Through Processing of auto claims. At First Notice of Loss, customers photograph vehicle damage via a smartphone app link, and Tractable's AI assesses the damage to build an estimate within minutes. In 2021, Admiral Seguros processed 12,000 touchless claims this way: 90% of estimates were processed without appraiser input, 98% of qualifying claims were completed in under 15 minutes, the AI's accuracy was 96% compared to human appraisers, and automation drove a 3x return on investment.
Pairing with AI for Public Sector Impact in Singapore
Since mid-2023, Open Government Products (OGP), an experimental tech team within the Singapore government, has rolled out the Pair suite of AI-powered tools for public officers: Pair Chat (a secure ChatGPT-like assistant), Pair Noms (meeting-minutes generation) and Pair Search (search over public government records). The products have over 60,000 registered users, more than 20,000 weekly active users, and over 10 million messages exchanged, saving officers an estimated 46% of time spent on administrative tasks. Data is kept private within government infrastructure and not logged on cloud providers.
Enterprise Data Analytics Platform and AMI Operations
Con Edison rolled out the C3 AI Platform and C3 AI AMI Operations application to establish AMI operations across 5 million smart meters and build a federated enterprise data analytics platform. The team aggregated two years of data from 13 source systems covering 5 million customer accounts, configured two machine learning algorithms and 50 analytics to identify deployment and installation issues and determine meter and network health. Results included $854M+ in identified annual customer benefit and 2,300+ deployment issues identified within 4 months.
ElektroDistribucija Srbije modernizes its grid with Schneider Electric's EcoStruxure ADMS and DERMS
ElektroDistribucija Srbije (EDS), Serbia's distribution system operator serving 3.8 million customers, implemented EcoStruxure ADMS and EcoStruxure DERMS from Schneider Electric to digitize grid operations, support renewable energy integration and improve reliability across its legacy electricity grid. Results included a 10-15% reduction in network losses, a roughly 20% reduction in outages, improved integration of distributed renewable energy resources, and condition-based maintenance improving operational efficiency.
Fiserv embeds Stuut Technologies' agentic AI into order-to-cash workflows, processing over $2 billion in invoices
Fiserv partnered with Stuut Technologies to integrate agentic AI into enterprise order-to-cash workflows through its Commerce Hub and SnapPay products. The initiative has already processed over $2 billion in invoices.
Hyundai Motor Group deploys autonomous mobile robots across its Metaplant America manufacturing process
Hyundai Motor Group's Autonomous Mobile Robot (AMR), which uses SLAM technology with LiDAR sensors and 3D depth cameras to navigate and select optimal travel paths, is currently being used across the entire manufacturing process at Hyundai Motor Group Metaplant America (HMGMA). The Group also showcased other AI robotics products at CES 2026 including the Atlas humanoid robot, Spot integrated with Orbit AI software for facility inspection, and the AI Keeper quality-inspection solution based on Spot that autonomously detects assembly defects on the automotive production line.
AVIS deploys Nativa-built reservation bots on LivePerson to automate car rental bookings
AVIS partnered with Nativa, which used LivePerson's conversational technology to build reservation bots that complete the car rental process in real time without human assistance, including quotes, confirmations, modifications and payments, via WhatsApp and Web Messaging. Through ongoing optimization from October 2021, bot understanding (Intent Match Rate) reached 88.3%, containment rate reached 69.9% (up from 8.4%), and average monthly customer satisfaction reached 91.4% in 2025.
Aurora Public Schools sees 28% improvement in literacy grade-level expectations using MagicSchool
Aurora Public Schools, a Colorado district of over 38,000 students, used MagicSchool's Enterprise package and MagicStudent tools including the Language Learning Tutor, Translate It! and Writing Feedback Tool to support English language learners, struggling readers and students with IEPs. A teacher, Johnnie Lacey, used the tools to differentiate instruction, and the district saw a 28% improvement in students meeting grade-level expectations for literacy, with the most significant gains among initially struggling students.
ADNOC generates $500 million in value from AI deployment across its value chain in 2023
ADNOC integrated over 30 AI tools across its full value chain, from field operations to corporate decision-making, generating $500 million (AED1.84 billion) in value in 2023 and abating up to 1 million tonnes of CO2 emissions between 2022 and 2023. Tools include the CPAD predictive analytics and diagnostics program, and AIQ-developed products Emission X, SMARTi, AR360 and Robowell.
How Schibsted is using AI to boost efficiency for their newsrooms and their readers
Schibsted Media (Norway) uses generative AI across its newsrooms via several tools: a suggested-stories feature at VG that maintains an 85% thumbs-up ratio comparable to lead stories; Videofy, an in-house service that generates 20-second news videos instantly from published articles, integrated into its Content Management System; and ARIA, an AI framework of specialised agents that lets staff query real-time compiled data through platforms like Slack. Journalists also use an AI toolbox, built by a journalist rather than the AI team, to generate potential questions and answers around articles using models including Claude and GPT.
Georgia State University tracks 800+ risk factors daily to improve graduation rates with predictive analytics
Georgia State University's predictive student retention analytics implementation tracks over 800 risk factors for more than 40,000 students daily and generates 90,000 targeted interventions annually. The approach has resulted in a 7-percentage-point increase in four-year graduation rates while closing achievement gaps among different student populations.
Building an Agentic Workforce: What We've Learned from 30,000 AI Agents
Prosus is building an agentic workforce across the group, with over 26,000 AI agents built as of December 2025 toward a target of 30,000 by end of March. At portfolio company iFood (Brazil), a Restaurant Account Executive Agent used daily by over 200 associates automates restaurant performance reporting, performing manual work equivalent to 40 full-time employees. A Data Analyst Agent lets employees query data in natural language without technical expertise. Prosus shifted from abstract workshops to hands-on sessions embedding the AI team into functions like communications, HR, sales, finance and legal to drive adoption, and tied agent creation to performance reviews and bonuses. Across the group, Prosus estimates it has created the equivalent of 1,000 full-time employees in capacity.
Dutch bank ABN Amro to cut 5,200 jobs by 2028 in AI-driven overhaul
ABN Amro announced plans to cut 5,200 full-time positions by 2028 (about 24% of its workforce) as part of a restructuring that relies on AI to handle routine back-office tasks. CEO Marguerite Bérard said AI would enable 'deliberate decisions' in customer service, operations and anti-money laundering checks, where headcount could shrink by up to 35%. The plan targets a return on equity of at least 12% and a cost-to-income ratio below 55% by 2028.
How Business Insider's AI-based paywall strategy increased conversions by 75%
Business Insider replaced its editorially-driven freemium paywall with a machine-learning-based smart paywall that decides which content to paywall for each user based on prior reading habits, referral platform, and each content genre's propensity to convert. In testing from December to April, total conversions increased more than 75% versus the control, with 60% of new conversions coming from stories that would never have been paywalled before. Applying AI to the registration wall as well increased registrations by 300%.
AIG uses generative AI to speed underwriting
AIG is using large language model technology to modernize underwriting as part of its AIG Next enterprise transformation. CEO Peter Zaffino said early pilots improved data collection and accuracy rates in underwriting from near 75% to upwards of 90%, while reducing processing time significantly. AIG has invested approximately $300 million in data, digital workflow, AI and talent over two years, and more than $1 billion on foundational data technologies over the last half decade, and plans to open a new data engineering and AI operations facility in Atlanta in 2026.
Apollo Global Management builds AI capabilities into its private equity portfolio companies
Apollo Global Management runs a structured playbook, led by its Data, Digital and AI team, to build AI capabilities into portfolio companies before and after acquisition. Cited results include: Cengage running eight AI projects with costs down 40% in select content production and a Student Assistant tutor bot scaled to 1 million users; Yahoo seeing engineering productivity gains exceeding 20% with over 10,000 lines of AI-generated code accepted daily; Barnes Group achieving a five-times return on a generative-AI documentation tool in its first year; and Univar Solutions' AI sales agent achieving a 30% engagement rate reactivating dormant customers.
Revolutionizing Audi's Welding Inspection System through AI
Audi implemented an AI-powered quality control system to transform its welding inspection process, expanding on previous manual ultrasound inspections that covered only 5,000 points per vehicle. The AI system analyzes 1.5 million weld points across 300 vehicles per shift, letting production staff focus on addressing potential anomalies rather than manual inspection. Following its success, Volkswagen Group is expanding the technology across multiple sites, including Audi Brussels, VW Emden and Audi Ingolstadt, with site-specific AI model retraining to accommodate varying welding parameters.
Westat builds machine learning models to triage SAMHSA's Drug Abuse Warning Network emergency department data
For SAMHSA's Drug Abuse Warning Network (DAWN) study, which collects data from 50 hospitals across the United States, Westat developed ML models to review and route emergency department records, assigning a probability score for whether a visit is in scope for DAWN and its likely category. The models are retrained periodically, producing high-quality DAWN data without relying on human review of every case.
Vanguard estimates AI ROI at close to $500 million across call center, advisor and coding use cases
Vanguard Group estimates its return on investment from AI at close to $500 million. Columnists Thomas H. Davenport and Randy Bean, writing in MIT Sloan Management Review, describe several proven AI use cases at the asset manager: AI agents in the call center help representatives draw answers from internal content and resolve issues faster; autogenerated, personalized summaries of Vanguard market perspectives help advisers keep clients informed; AI-assisted code generation has improved programming productivity by 25% and trimmed the system development life cycle by as much as 15%; and a large language model analyzes corporate earnings calls for signals that dividends will be cut. Vanguard also continues to monitor AI model performance and utilization, and reports that 50% of employees have completed training through the Vanguard AI Academy.
Target rolls out GenAI-powered Store Companion chatbot to stores chainwide
Target announced plans to roll out Store Companion, a generative AI chatbot built by its in-house technology team, to team members at all of its nearly 2,000 stores by August 2024. The tool is designed to answer store team members' on-the-job process questions, coach new hires and support store operations. As of the June 2024 announcement, Target was piloting the tool at about 400 stores, using pilot-team feedback to improve the experience ahead of the planned chainwide rollout, following an initial testing phase to planned rollout in just six months.
UK Solicitors Regulation Authority Approves Garfield.Law as First AI-Driven Law Firm
In May 2025 the UK Solicitors Regulation Authority (SRA) authorised Garfield.Law Ltd, co-founded by Philip Young and Daniel Long, as the first law firm authorised to deliver legal services entirely through AI technology. Garfield operates an AI-powered litigation assistant for small claims debt recovery up to £10,000, guiding users from calculating amounts owed through filing for default judgment, and charges per document rather than hourly. The SRA authorisation included safeguards such as client confidentiality protection, conflict avoidance, user approval at each stage, and controls preventing the AI from proposing case law to address hallucination risk.
KI hilft bei Onlineanmeldung der Eheschließung
Die Landeshauptstadt Wiesbaden setzt bei der Onlineanmeldung zur Eheschließung nun Künstliche Intelligenz ein. Die KI generiert direkt nach der Anmeldung eine individuelle Übersicht der benötigten Dokumente für das Traupaar, analysiert die hochgeladenen Unterlagen anhand von zehn typischen Konstellationen, prüft Vollständigkeit und Lesbarkeit und bereitet die Dokumente für die fachliche Prüfung vor. Ein Ampelsystem zeigt den Standesbeamtinnen und Standesbeamten, ob Dokumente fehlen. Die rechtliche Prüfung und Entscheidung verbleiben beim Standesamt. In Wiesbaden melden 75 Prozent der Traupaare ihre Eheschließung online an.
Dutch National Archive uses AI to make its largest WWII archive digitally searchable
The Nationaal Archief, Netwerk Oorlogsbronnen, Huygens Instituut and NIOD Instituut are digitizing the Centraal Archief Bijzondere Rechtspleging (CABR), the largest and most-consulted Dutch WWII archive with dossiers on roughly 300,000 people, and using artificial intelligence to make the digitized documents readable and searchable. The first digital results go online via Oorlogvoorderechter.nl from 2025, with the project running until 2027.
Volvo Trucks uses proprietary AI models to power adaptive maintenance in its Blue Service Contract
Volvo Trucks North America upgraded its Blue Service Contract to offer optimized service stops based on fuel consumption, idle time and oil samples, powered by the OEM's proprietary AI models, enabling fewer stops for 2023 model year and newer trucks and reducing operating costs for customers.
Booking.com launches Smart Messenger and Auto-Reply agentic AI tools for partner-guest communication
Booking.com's first customer-facing agentic AI tools, Smart Messenger and Auto-Reply, gather partner, property and reservation information to autonomously suggest and send guest communications; early experiments yielded a 73% increase in partner satisfaction compared to previous messaging tools. The company also offers AI Trip Support, AI Voice Support, AI Rental Helper and AI-powered Flight Search Summaries.
Talkmobile: Setting the standard for modern, digital customer service
Talkmobile, a UK-based mobile network operator founded in 2007, has partnered with LivePerson since 2014 on a conversational AI strategy spanning web messaging, social and digital channels, secure forms, and an intelligent routing bot. LivePerson's conversation data analytics let Talkmobile automate common inquiries such as bill payments and account updates on its Ecare platform, while migrating from web chat to asynchronous web messaging in 2022 drove a 60% increase in average monthly volume within six months. Talkmobile now maintains a 92% customer satisfaction score, a Net Promoter Score of 66, and an 89% first contact resolution rate, with agents responding to customers in under one minute using LivePerson's Agent Workspace.
Strava's Athlete Intelligence Translates Workout Data into Simple and Personalized Insights
Strava launched Athlete Intelligence, an AI-powered feature available as a public beta to subscribers, which analyzes and interprets workout data across pace, heart rate, elevation, power, and Relative Effort into simple, personalized insights and guidance. The feature spots 30-day performance trends, detects milestones such as fastest pace or longest distance, and offers tailored feedback for each activity, drawing on more than 10 billion activity uploads on Strava.
Süddeutsche Zeitung: Lokalinformationen per WhatsApp
Süddeutsche Zeitung (SZ) launched WhatsApp channels for districts around Munich, offering free local news, event information and regional updates. Content comes from SZ's local newsrooms and is supplemented by automatically generated information produced in cooperation with Berlin media startup Beat Squares, which uses AI to process data from publicly accessible sources. All contributions are editorially reviewed before publication. More than 4,000 users subscribed within the first three days.
Deutsche Telekom partners with McKinsey to scale hyper-personalized agent training with gen AI
Deutsche Telekom partnered with McKinsey to build a generative-AI-powered learning and coaching engine to upskill 8,000 human agents in the field and call centers. The team spent six weeks diagnosing agent needs with millions of data points, then four months building and refining the MVP, embedding prompts into agent workflows (e.g. a training video is triggered if an agent struggles with eSIM activation). The likelihood of customers recommending the company increased by 14%.
Bankdata pilots Microsoft's agentic COBOL Migration Factory against its 70-million-line mainframe estate
Bankdata, a technology company owned by a consortium of Danish banks representing over 30% of the Danish banking market, has over 70 million lines of code still running on the mainframe. Bankdata worked with Microsoft to build the COBOL Agentic Migration Factory (CAMF), a multi-agent system built on Semantic Kernel with specialized agents (COBOLAnalyzerAgent, JavaConverterAgent, DependencyMapperAgent) that analyze, map dependencies for, and convert COBOL programs into modern Java Quarkus applications. Microsoft tested the framework on a small COBOL module donated by Bankdata, progressively validating it on more complex COBOL call chains reaching a call-chain depth of level 3 as the tool matured into its current multi-agent form.
Noxtua's sovereign European legal AI in use at Allianz, Deka and law firm Nordermann
Berlin-based Noxtua (formerly Xayn) has built what it says is Europe's first 'sovereign' legal AI, with large language models trained exclusively on European legal data and hosted on EU cloud providers. The tool, which helps lawyers research, analyse and draft documents, is already in use by law firms such as Nordermann, the legal department of Allianz, and German savings bank Deka. Noxtua raised an €80.7m Series B led by legal publisher C.H.Beck in April 2025.
BMW advances engineering simulation with Mistral
The BMW Group is partnering with Mistral AI to develop industry-specific AI models for crash simulation, combining BMW's proprietary datasets and engineering expertise with Mistral's AI capabilities to improve quality, accuracy and speed for complex engineering tasks. The collaboration focuses on embedding domain-specific knowledge into AI workflows, starting with crash simulation and with potential to scale across further applications along the BMW Group value chain.
ASML accelerates advanced semiconductor lithography with Mistral
ASML is partnering with Mistral AI to apply frontier AI across its product portfolio as well as research, development, and operations to accelerate time to market with higher performance holistic lithography systems. Initial use cases include AI agents for software reliability and early downstream issue detection, large-scale log analysis with intelligent reasoning for root-cause identification in diagnostics, and automation of technical product documentation within engineering workflows.
Capgemini pilots AI-RAN-connected autonomous shuttles under Horizon Europe's Project ULTIMO
Capgemini is working within Project ULTIMO, a Horizon Europe-funded initiative, to demonstrate AI-RAN support for large-scale autonomous mobility across European cities. Autonomous shuttles equipped with NVIDIA Jetson Orin modules process sensor data locally, while select video and telemetry streams are sent over 5G to agentic AI applications running on NVIDIA AI-RAN servers for scene understanding, incident and safety detection, and accessibility insights, with mission-critical 5G traffic given priority access to GPU resources.
Unveiling Magic Apron: The Home Depot's Smartest Tool Yet
The Home Depot expanded Magic Apron, its suite of generative AI tools that helps customers find answers to home improvement questions, available on The Home Depot app and on millions of product pages on homedepot.com. It is powered by a proprietary home improvement knowledge base combining large-scale datasets with The Home Depot's own expertise and product information, and already answers customer questions and summarizes product reviews. It is planned to expand to The Home Depot's Pro site for professional contractors and business account users.
Singapore General Hospital trials AI2D prescribing assistant for antibiotic dosing
SGH collaborated with DXC Technology and Synapxe to develop Augmented Intelligence in Infectious Diseases (AI2D), an AI-based prescribing assistant mainly used for the diagnosis of patients with pneumonia and urinary tract infections. AI2D lets doctors know whether antibiotics are needed and recommends the type and dosage of antibiotics based on the patient's clinical data.
Rentokil Initial deploys Gemini-powered assistant for field technicians with Telana
Rentokil Initial partnered with Telana to deploy a Gemini-powered assistant that provides field technicians with hands-free, real-time service and customer data.
Mia: Houston Methodist's intelligent chatbot for nursing recruitment
Houston Methodist's Center for Innovation built Mia, an intelligent chatbot that interfaces with nursing candidates early in the recruitment process to vet them for appropriate interest and experience.
McGee & Co uses Syte's visual AI to connect shoppers with complementary home decor pieces
Furniture and home decor retailer McGee & Co implemented Syte's visual search and AI tagging technology to tie products together, helping customers who find an item like an end table or couch also discover the complementary pieces from the same collection. Josh Batchelor, VP of Technology at McGee & Co, said the goal was to make sure shoppers are served complementary pieces when they need them.
Frankfurter Allgemeine Zeitung relaunches app "FAZ. Der Tag" with AI-generated article summaries
Frankfurter Allgemeine Zeitung redesigned its news app "FAZ. Der Tag," which presents the most important topics of the last 24 hours along with a daily podcast and news quiz. The relaunched app offers AI-generated article summaries for a quick overview, alongside personalization by topic, a new design with improved navigation, and a read-aloud function. The publisher introduced a new subscription model for the app: free for the first 30 days, then €0.99 for the first month and €7.99 per month thereafter, with free access for FAZ+ and FAZ/FAS digital subscribers.