AI News NQ Analysis

Chiyoda X-One Engineering Implements AI x RAG Technology to Systematize Internal Knowledge with 'Lightblue' AI Agent

NQ Score 96/100

AI Summary (NQ-processed)

Chiyoda X-One Engineering has begun full-scale operation of the 'Lightblue' AI agent to leverage its vast internal knowledge base using RAG technology. This initiative aims to improve operational efficiency, train junior employees, and enhance design quality by making tacit knowledge explicit and accessible.

AI Analysis

Frequently Asked Questions

Q: Why did Chiyoda X-One Engineering introduce the AI agent 'Lightblue'?
A: To effectively utilize its vast, underused internal knowledge assets, resolve issues of knowledge siloing, accelerate the training of junior employees and technology transfer, and improve the efficiency and quality of design and engineering work.
Q: What key feature of 'Lightblue' was crucial for its adoption?
A: Its RAG (Retrieval-Augmented Generation) function, which allows the AI to reference internal company databases to provide accurate, context-specific answers, along with its high answer accuracy and ease of implementation.
Q: How does the company ensure data security while using this AI service?
A: 'Lightblue' provides robust security measures, including a setting where input data is not used for model training and operation in an environment isolated from external networks. Chiyoda X-One Engineering also conducted a vendor security check to confirm it meets their required security levels.
Q: Which departments are involved in the rollout of 'Lightblue'?
A: The initial trial was led by the Power & Control Project Division. The full-scale operation involves distributing IDs to multiple divisions, including Engineering, Energy & Environment Projects, Life Science Projects, Site Operations, and Corporate Management, with a goal of eventual company-wide deployment.
Q: What is 'Lightblue'?
A: It is a next-generation AI agent service for corporations and municipalities developed by Lightblue Inc., a startup from the University of Tokyo. It can perform tasks like search, summarization, and transcription, and allows users to build custom, secure chatbots without coding.