United StatesManufacturing2025 – ongoingToyota Motor North America
Multimodal LLM on repair manuals and trouble tickets for process-engineer troubleshooting
Toyota North America's AI chief says a calibration-reset task fell from 7 hours to about 15 seconds with a multimodal LLM
Problem
Process engineers had to troubleshoot line-level machinery such as calibration resets, which involves G-code and large manuals that cannot be searched easily. A single issue could take hours.
Approach
Toyota Motor North America fed repair manuals, trouble tickets and video documentation, some newly digitized, into a multimodal LLM. The result was a set of knowledge tools that cut diagnosis time and gave newer employees access to veteran-level know-how. The company also used AI coding tools for developers.
Results
Takeaway
The large time cut was reported for one calibration-reset task, so it is a single example. The developer gain comes from an earlier study.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
An interview with Brian Kursar, head of enterprise AI at Toyota Motor North America. He described a multimodal LLM built on manuals, trouble tickets and video, and said one troubleshooting task dropped from seven hours to 15 seconds. He also said developer productivity rose by at least 20%, based on an earlier study.
Full details are in the original sources.