Where the AI runs
On-premise, hybrid or public cloud — recorded per entry. The private-AI axis no hyperscaler directory keeps.
Source-verified register · updated 2 Sept 2026
Proven Use Case Library for Enterprise AI.
Every entry records where the AI actually runs, which regulations touch it, and the source behind the claim — with every figure checked against that source before it publishes.
Unverifiable entries are unpublished, not softened — takedowns land on the register's own ledger.
Register signalslive
Where the AI runs
Regulations tagged in the register
100%of published entries carry a primary source
- Named customer335/335
- Primary source335/335
- Quantified outcome324/335
335 published entriesOpen the register
335published use cases, each linked to its source
28industries covered, from healthcare to logistics
3deployment models recorded per entry, not guessed
tagged per entry, alongside NIS2, DORA and ISO 42001
Latest additions
The newest entries on the register. Each links to the source behind its claims.
Last added
This week+42entries added
This month+335entries added
Deutsche Telekom and Google Cloud Collaborate on MINDR Agentic AI for Autonomous Network Operations
Deutsche Telekom, in partnership with Google Cloud, announced MINDR (Multi-Agentic Intelligent Network Diagnostics & Remediation), a multi-agentic AI system built with Google Gemini models on Vertex AI and developed with Google Cloud's Autonomous Network Operations framework. MINDR correlates signals end-to-end across the RAN, transport, and core network domains to proactively identify service-impacting issues and support autonomous, explainable remediation, shifting operations from reactive troubleshooting to predictive automation. MINDR builds on Deutsche Telekom's RAN Guardian Agent, live in production in Germany since November 2025, which monitors network behavior during high-impact events and traffic surges. RAN Guardian autonomously triggered over 100 remediation actions at Christmas market events in its first month and reduced the time needed to manage major events from hours to around a minute (a more than 95% improvement). For 2026 it identified 237,000 events, including around 130 Carnival events each expected to draw over 10,000 participants, served by 611 mobile sites pre-checked for issues; only 5 sites experienced peak loads, which RAN Guardian optimized during the event. RAN Guardian is now scaling across Deutsche Telekom's European national companies, starting with the Czech Republic and Croatia. MINDR's first production releases are planned for later in 2026.
96/100High6 of 7 signalsPrimary source on recordAdded
Deutsche Telekom· GermanyGoogle Gemini · Google Vertex AI
NatWest Group Takes on Digital Transformation to Make Home Buying Easier
NatWest Group, a UK banking and financial services company headquartered in Edinburgh serving approximately 19 million people, families, and businesses throughout the UK and Ireland, co-created an AI-powered, cloud-based platform called "Marge" with IBM Consulting. Built with IBM watsonx Assistant technology on IBM Cloud, Marge gives mortgage call center employees real-time digital mortgage support, embedded in NatWest's existing data structures with new data added every minute through content updates and customer interactions. Employees access mortgage policy information by typing keywords into a console while on calls with customers. Since implementing the tool, NatWest saw a 20% improvement in customer Net Promoter Score (a customer loyalty metric) and a 10% decrease in call duration.
96/100High6 of 7 signalsPrimary source on recordAdded
NatWest Group· United KingdomIBM watsonx Assistant · IBM Cloud
Galene.AI Deploys a Sovereign-by-Design Conversational AI Platform on Scaleway
Galene.AI built a fully sovereign, self-hosted conversational AI platform running on Scaleway's GPU infrastructure, engineered so client data, models, and compliance remain entirely within European jurisdiction. The platform deploys as a private Kubernetes cluster, either on-premise or within a private Scaleway cloud, with all AI inference processed locally on Scaleway GPUs (from 2xL40S to H100-based configurations) rather than through external API calls. Data is encrypted in transit (TLS 1.3) and at rest (AES-256). The platform includes an agentic framework using Model Context Protocol (MCP) to connect private AI agents to internal knowledge bases and enterprise systems (CRMs, ERPs), and a built-in governance layer called Generative Shield that enforces real-time compliance across cybersecurity, ethical/legal, data protection, and output-accuracy dimensions. Scaleway's data centers are located exclusively in the EU (France, Poland, Netherlands), which the companies say eases GDPR data-residency compliance and EU AI Act compliance. Galene.AI's REST API is compatible with OpenAI's API specification.
100/100High7 of 7 signalsPrimary source on recordAdded
Galene.AIKubernetes · Model Context Protocol (MCP)
WithSecure Built Luminen, an AI Cybersecurity Assistant, on Amazon Bedrock
Cybersecurity company WithSecure built Luminen, a generative AI assistant on Amazon Bedrock, to help organizations detect and respond to security events faster, minimizing damage and reducing downtime. WithSecure provides cybersecurity solutions for mid-market companies, prioritizing privacy, data sovereignty, and regulatory compliance in the European Union. Luminen integrates with WithSecure Elements and includes a Security Awareness Assistant that analyzes complex data tables to provide a 7-day security status overview, and an Investigation Assistant that analyzes technical data on identified security issues to build a narrative. Development began in December 2023, a public beta released in May 2024, and Luminen launched in September 2024. WithSecure also uses Amazon Neptune to build knowledge graphs storing relationships between security events.
96/100High6 of 7 signalsPrimary source on recordAdded
WithSecureAmazon Bedrock · Amazon Neptune
myTomorrows Matches Patients to Clinical Trials Using AI Hosted in Europe
myTomorrows is a health tech platform that uses AI to match deidentified patient medical profiles against clinical trials and expanded access treatments not yet approved for public use. AI-based matching saves around 90 percent of the time compared to manual review by a medical professional, turning a two-hour task into about one minute. Because the platform processes health data, its large language models are hosted in Europe on AWS to comply with GDPR.
100/100High7 of 7 signalsPrimary source on recordAdded
myTomorrowsAmazon Web Services (AWS)
Coverage by sector
Where the register is deepest today. Every count is the live published set, nothing projected.
Financial Services48
Manufacturing29
Automotive25
Technology & Software24
Healthcare21
Retail21
Government & Public Sector21
Energy18
Published entries only.
What a hyperscaler directory will not tell you
Four fields we record on every entry. They are the reason a European buyer can shortlist from this register instead of starting a research project.
Which regulations apply
Every entry carries the frameworks that touch it, so a compliance conversation starts from a list instead of a blank page.
How strong the evidence is
Every entry carries a computed 0–100 score in which each point is a named signal on the record. High evidence is unreachable without a primary source and a named customer, and low-evidence entries are hidden until you ask for them.
HighMediumLow
Checked against the source
Every checkable claim is tested against the fetched source, and one unsupported figure fails the whole entry. Nothing publishes on a template, and anything whose source later contradicts it comes down rather than getting softened.
Source-verifiedPublish gate: a passing verification, still matching the entry's content
Shortlist from evidence, not from vendor decks.
Free to browse. Filter by sector, capability and deployment model, and open the source behind any claim.