From Knowledge Held by One Person to Knowledge AI Can Draw On

Local government work depends heavily on the tacit knowledge of individual staff. This knowledge is not an organizational asset and is always at risk of being lost when people transfer or retire.

In this studio, participants learn together while structuring tacit knowledge from their work and turning it into organizational knowledge that AI can understand and use. The key is not introducing AI tools but preparing context that AI can understand. We call this approach knowledge engineering. Local governments can then use AI on their own terms, rather than being used by it.

Common Challenges

• Generative AI tools have been introduced, but only some staff use them
• Judgments and procedures that staff follow implicitly are not shared
• Work stops when a particular person is absent
• Handovers take a long time because know-how is not documented
• Requests arrive by Excel and email and are re-entered by hand

What the Studio Aims For

• Less dependence on individuals and more continuity, by structuring judgment criteria and past cases as shared assets
• 30-50% less manual work in Phase 1, by replacing Excel-and-email requests with unified forms
• A data foundation that AI can use, structured as context (Phase 2)
• An AI adviser for each department that draws on accumulated knowledge to support decisions (Phase 3)
• Benefiting from AI while keeping control of the data that must be protected

Approach

Based on Code for Japan's experience supporting local governments, we work through these steps:

1. Put the goal into words
2. Review the current situation
3. Choose the target work and department
4. Analyze the work and decide where to use AI
5. Run a pilot and measure the results
6. Review and plan for implementation and expansion to other areas

Participating Local Governments

More than 10 local governments take part, centered on a monthly online study session. Each applies knowledge engineering to its own work, with Code for Japan members providing hands-on support on each theme, and participants share cases and challenges to learn from each other. Applications from local governments, companies and individuals are currently closed. Any additional call will be announced in our activity reports.

Activity Reports (in Japanese)

• Pre-kickoff event (March 2026)
• Kickoff event (April 2026)
• How AI-ready local governments work (September 2026)

Contact

If you are interested in knowledge engineering or hands-on support for AI use in local government, please contact us through our contact form.