Your External Brain: methodology overview
The thing you can do but can't explain
You've worked in your field long enough that the answers come automatically. A client asks why you do something a certain way and you can answer in real time, but the moment they ask you to write it down, you realize the answer that came out of your mouth wasn't a sentence you'd ever actually composed. It was assembled, on the fly, from years of compressed experience.
That's tacit knowledge. It's the most valuable thing you own and the easiest thing to lose.
Why a blinking cursor doesn't help
If you've ever tried to write down "how you do what you do," you know the trap: a blank page asks the wrong question. Tell me your method assumes you have the method ready. If you did, you'd have written it years ago. The reason you haven't isn't laziness; it's that the method isn't sitting somewhere accessible. It has to be reconstructed, and reconstructing it requires the right kind of attention from the right kind of interlocutor.
Most note-taking apps and AI chats give you a blinking cursor and wait. That's not enough.
What the experience looks like
When you start a session, the assistant doesn't ask you to dump your knowledge. It asks you what you want to systematize, what objective you have, and what kind of result you want to walk away with: a methodology document, a diagram, a template, an article.
From there it moves through guided phases. You'll be asked indirect questions, never "explain your method," because direct questions only retrieve answers you already had ready. The point is to surface what you didn't have ready. Some of the conversation is gathering material; some is the assistant proposing hypotheses for you to confirm or correct. Anything pulled in from the web or other tools is marked as unverified until you weigh in.
You can pause at any point. The work-in-progress is saved continuously, and when you come back the assistant resumes from exactly where you stopped, including which pieces of information you'd confirmed and which were still pending.
What you walk away with
Two artifacts. The final result itself: a real document, in the format you specified, that you can share, publish, version, or hand to a junior colleague. And a systematization document: the audit trail of how the result was built, including the hypotheses, the questions, and the changes in your own thinking along the way. The deliverable is what you publish; the audit trail is the receipt that proves how you got there.
Over time, a third thing accumulates: a library of reasoning blocks. Each completed systematization is indexed into atomic, structured pieces of your expertise: a method, a principle, a decision pattern, organized by how you think. These blocks stay available to your AI across future sessions, so the next time you ask it to reason about something in your field, it can draw on what you've already systematized instead of starting from scratch.
Why this works
The structure is invisible to you while you're working, but it's why the conversation feels different. Indirect questions retrieve material you didn't know you had. The verification rules keep unconfirmed information from quietly shaping your conclusions. The guided phases keep the conversation from wandering. You're not just talking to an AI; you're being interviewed by something that knows when to ask, when to wait, and when to write.
How to start
You need three things: an alpha invite, an AI assistant that supports MCP, and something you want to systematize. The connection takes under five minutes; there's a walkthrough at Connect your AI assistant to your External Brain. Once you're in, type something like:
"Help me systematize my thinking about [a method, a recurring decision, the questions you always ask before starting a project]."
The assistant will take it from there.