← Back to ARUM Adds an LLM-Powered AI Character to Machining Centers Using Azure AI

Source proof for ARUM Adds an LLM-Powered AI Character to Machining Centers Using Azure AI

Source-bound proof

Verified source excerpts for every supported field

Each colour maps a published value to the exact source passage used to support it. Only bounded excerpts are public; administrators can inspect the complete captured source.

13 fields supported

Title

ARUM Adds an LLM-Powered AI Character to Machining Centers Using Azure AI

derived · high
ARUM tackles skilled labor shortages by adding LLM to machining centers with Azure AI | Microsoft Customer Stories This is the Trace Id: 0476cb2fdec75143a53cc6a07c5a…

Description

Japanese manufacturing automation company ARUM developed TTMC Origin, a machining center operated through natural conversation with an LLM-powered AI character named KAYA, using Azure Speech, Azure OpenAI and Azure AI Search with GPT-5 running on an industrial PC. Users with no machining knowledge can perform high-precision operations by voice chat. Automating NC programming cuts the steps with CAM from 177 to 2, and since NC programming accounts for 50% of production cost per part, ARUM says automating it could cut production costs in half.

derived · high
…ning knowledge can perform high-precision operations via voice chats with KAYA. The tool’s Azure AI services include Azure Speech for voice communication with users, Azure OpenAI for summarizing and interpreting conversations, and Azure AI Search for database searches. Additionally, Foundry Agent Service on the Microsoft Foundry platform provides…
…nvolves 177 steps, but TTMC Brain automates all but the first and last stages. “NC programming accounts for 50% of the production cost per part, so automating this stage could cut production costs in half.” Ease of use is another essential point. Because users converse with the AI cha…

Company

ARUM

quote · high
…LLMs offer a potential solution to skill shortages in the metalworking industry Established in 2006, ARUM is transforming manufacturing with automation, robotization, and AI. In 2021, the company launched its manufacturing-focused ARUMCODE AI model to au…

Country

Japan

quote · high
…ND SUPPORT Microsoft for Startups Founders Hub ORGANIZATION SIZE 1-49 employees COUNTRY/REGION Japan BUSINESS NEED Customer experience Employee experience Artificial Intelligence I…

Industry

Manufacturing

classification · high
…n BUSINESS NEED Customer experience Employee experience Artificial Intelligence INDUSTRY Discrete Manufacturing Share the story Explore related stories /h2> "> Explore related stories Stories…

Problem

Japan has over 200,000 metalworking companies, but the number of skilled workers in the industry is dwindling drastically. ARUM had automated NC program generation and machining processes, but the many buttons on its tools still hindered effective use among inexperienced users.

derived · high
…ours were the first to automate NC program generation and machining processes.” However, the many buttons on these tools still hindered effective use among inexperienced users. ARUM therefore began considering large language models (LLMs). Its aim was to c…
…ditionally, creating and operating NC programs has required veteran expertise. “Japan has over 200,000 metalworking companies, but the number of skilled workers in the industry is dwindling drastically; we do not expect to see more in the future,” says ARUM CEO Takayuki Hirayama.…

Solution

ARUM developed TTMC Origin, a machining center operated through natural conversation with the LLM-powered AI character KAYA. An LLM (GPT-5) running on an industrial PC in the machining center controls KAYA, which integrates with Azure services: Azure Speech for voice communication, Azure OpenAI for summarizing and interpreting conversations, and Azure AI Search for database searches. Foundry Agent Service on the Microsoft Foundry platform provides a fallback generic AI in case the primary AI function does not perform as expected. KAYA also integrates with TTMC Brain, which, based on the conversation, automatically generates a program for the machining process and autonomously controls the machining center.

derived · high
…oasts a large vertical display, on which KAYA converses with the user by voice. An LLM (GPT-5) running on an industrial PC in the machining center controls KAYA, which also integrates with various AI services on Azure. Users with no machini…
…AI function doesn’t perform as expected. KAYA also integrates with TTMC Brain. Based on the conversation with the user, TTMC Brain automatically generates a program for the machining process and autonomously controls the machining center. Satoshi Murakami, General Manager of the ARUM CSE Development Team, explains th…

Implementation approach

CEO Takayuki Hirayama proposed the concept to Microsoft in 2023. ARUM then started developing TTMC Origin with the support of Microsoft for Startups, which supported the project from concepting to prototyping, MVP design and bringing the solution to market. A prototype was completed in August 2025.

derived · high
…d by 50% Microsoft supports development and implementation of AI character KAYA After the proposal, ARUM started developing TTMC Origin with the support of Microsoft for Startups. This program propelled everything from concepting to prototyping, MVP design, a…
…rom concepting to prototyping, MVP design, and bringing the solution to market. A prototype was completed in August 2025. The machining center supports simultaneous three-axis high-speed rotation and f…

Business value

TTMC Origin decreased NC programming steps with CAM from 177 to 2. Because NC programming accounts for 50% of the production cost per part, ARUM's General Manager says automating this stage could cut production costs in half. ARUM says reservations open in July 2026 with inquiries already flooding in, and it plans to release the solution overseas gradually from 2027.

derived · high
…ant to elevate the global manufacturing industry.” Takayuki Hirayama, CEO, ARUM NC programming steps with CAM decreased from 177 to 2 Manufacturing cost per part reduced by 50% Microsoft supports development and i…
…AM. Users can perform high-precision machining as easily as using a smartphone. Reservations open in July 2026 and we’re already flooded with inquiries. We’ll gradually release it overseas from 2027. We want to elevate the global manufacturing industry.” Takayuki Hirayama, CEO,…

AI capabilities

Conversational AI, Generative AI

classification · high
…. The machining center supports simultaneous three-axis high-speed rotation and features TTMC Brain and LLM-powered AI character KAYA for chat. The front of the machining center boasts a large vertical display, on which KA…

Technology

Azure AI Search, Azure Speech, Azure OpenAI, Microsoft Foundry, Foundry Agent Service, GPT-5

classification · high
…t have embarked on will be unmissable. Discover more details CUSTOMER ARUM Inc. PRODUCTS Azure AI Search Azure Speech in Foundry Tools Azure OpenAI in Foundry Models Microsoft Foundry Foundry Agent Service SERVICES AND SUPPORT Microsoft for Startups Founders Hub ORGANIZATION SIZE 1-49…

Headline outcome

derived · high
…nvolves 177 steps, but TTMC Brain automates all but the first and last stages. “NC programming accounts for 50% of the production cost per part, so automating this stage could cut production costs in half.” Ease of use is another essential point. Because users converse with the AI cha…

Use case type

Autonomous systems

classification · medium
…KAYA also integrates with TTMC Brain. Based on the conversation with the user, TTMC Brain automatically generates a program for the machining process and autonomously controls the machining center. Satoshi Murakami, General Manager of the ARUM CSE Development Team, explains th…
Capture details
Captured
12 Oct 2026, 05:16 UTC
Extractor
fetch-strip@1
Snapshot hash
131eebb62a30460920948db3924451aa1db4b7c2229658d5b2947a442656f9d1