Sources and further reading
The cases, figures and rules the article draws on, by chapter, with what each one supports. The rail maintenance company followed throughout the article is a composite.
01 · Anatomy of a RAG
Toyota is deploying AI agents to harness the collective wisdom of engineers and innovate faster
About 800 powertrain engineers using a system built so that retiring experts' knowledge is not lost, drawing on design reports, regulations and veteran engineers' handwritten notes, with drawings left for later.
Lack of AI-Ready Data Puts AI Projects at Risk
The prediction that organisations will abandon 60% of AI projects unsupported by AI-ready data, and the 63% of organisations that lack the right data practices or do not know whether they have them.
03 · Governance first, kept light
Regulation (EU) 2024/1689, the Artificial Intelligence Act
Article 5 lists the prohibited practices, including social scoring and emotion recognition in the workplace. Article 4 requires an adequate level of AI literacy among staff who use AI.
04 · What does “better” mean?
An assistant drawing on a corpus of 100,000 documents, every use case tested against an evaluation framework before deployment, daily regression testing, and adoption by more than 98% of advisor teams.
Elsewhere on this site
GenAI Beyond the Hype, chapter 04
The crawl, walk, run, fly stages used in chapter 05, and the governance pillars behind chapter 03.
The Definition of Done, timeboxes and planning under uncertainty, used in chapter 06.
Scrum as a framework for work whose outcome must be discovered step by step, and the product owner's role, used in chapters 02 and 07.