Thinking tool: Be AI-legible

This is my personal blog, where I write solely in my personal capacity. It does not represent the positions of any organisations I'm associated with.

Also, everything here is provded "as is", without warranties of any kind, express or implied.

18 August 202618 August 2026

AI systems, like Claude, are extremely powerful force multipliers for people who know how to use them well.

This force multiplication scales with how much AI is actually able to find out about your work and take actions in the world on your behalf.

Two panels, each with an AI chat bubble (two sparkles inside) casting a light cone over documents. Left: a narrow beam catches one document of five, the rest sit in the dark - AI sees a sliver of your work. Right: all five documents sit inside a wide beam - AI sees all your work

You should therefore make your work AI-legible. Some behaviours that contribute to this right now might include:

Have a good connectors setup
Auto-transcribe all your meetings (you can do this with Google Meet)
Keep written notes somewhere accessible, e.g. in Google Docs (+ have the Google Docs MCP!) rather than a physical notebook
If you whiteboard something, take a photo and post it somewhere indexable
Write a top-level system prompt that explains which information sources exist and where to find things. e.g. that you use Notion for some docs and Google Drive for others, or that certain information is stored in GitHub, some in issues/PRs/wikis and some in the repository itself.

This is part of my thinking tools series. Also consider:

Become a narrow expert: works much better when the AI can see your actual context

Thinking tool: Be AI-legible

Related concepts