How to Build an AI Second Brain Your Tools Can Actually Read (The Weekend Setup)
Build a second brain your AI can actually read: the folder structure, the README template, and an honest GitHub vs Obsidian vs Notion vs Drive comparison.
Everything about how I run my agency lives in one folder of plain files: every meeting note, every strategy doc, every content plan, every client's history, and all the rules my AI runs on. People call this a second brain. Mine runs my company, and this is the setup manual for building your own in a weekend.
Step 1: Pick the home
Your second brain needs one home, and the right one is whichever you'll actually maintain. The four worth considering:
- GitHub is what I use. Built for plain files, keeps a full history of every change, free, and AI tools read it cleanly. It's also the most technical on-ramp of the four. I pay that cost happily for total portability, but for most people it's overkill, and deciding that is fine.
- Obsidian keeps plain markdown files on your own computer. Fast, offline, great for a solo brain. Syncing across a team takes extra setup.
- Notion is the most approachable and the most collaborative, and the most locked-in. Your content lives in Notion's format, and getting it back out cleanly is harder. Strong pick if your team already lives there; go in knowing the tradeoff.
- Google Drive is the one you already have. Familiar and collaborative, though native Google Docs read a little less cleanly for AI, and Drive folders sprawl without discipline.
One non-negotiable, whatever you pick: your AI needs a way to read it, usually through MCP (Model Context Protocol, the standard plug that lets an AI tool read another app's files). A second brain your AI can't open is just a filing cabinet.
Step 2: Set the structure
Structure is what separates a second brain from a junk drawer.
- One top-level folder per client or project. (Personal version: one folder per area of life. Health, home, money, travel.)
- The same subfolders inside every one. Meeting notes. Intake and onboarding. Strategy. Content plans. Copy. Data. Same layout every time, so neither you nor the AI has to go hunting.
- Folders that mirror your org chart, if you're using AI as a team. I have a marketing folder, and inside it, the department: director of marketing, content marketer, social media manager, copywriter, graphic designer. Each role is its own instruction file.
Step 3: Put a README at the top of every folder
This is the piece most people are missing, and it's what makes the whole thing work. A README is a plain-text file that sits at the top of a folder and explains what's inside and how to use it. (The name is a programming convention. It literally means "read me first.") It's an onboarding doc for the tool.
Here's the skeleton I use. Copy it:
# [Client/Project Name] — README
## What this folder is
One paragraph: who this client is, what we do for them, and the
current state of the work.
## How we work with them
- Their preferences (formats, tone, pet peeves)
- Things they never want
- Their internal vocabulary (what they call things)
## Where things live
- /meeting-notes — every call, filed by date
- /strategy — positioning, brand guidelines, plans
- /content — drafts and shipped work, by month
- /data — analytics and performance
## Rules for working in this folder
- Read the two most recent meeting notes before drafting anything
- All copy follows the shared voice rules in [path to your rules doc]
- Flag anything that contradicts the strategy doc instead of guessing
When you ask for a deliverable, the AI reads the README, picks up the rules, and knows exactly where to look.
Step 4: Write the shared rules once
Your writing rules, your banned words, your repeatable how-tos live in ONE shared spot that every folder points back to. Write a rule one time, and everything downstream follows it. (Building that doc is its own guide: see "Write Your AI's Rules Doc.")
Step 5: Connect your AI and prove it reads
Connect the tool to your chosen home, then test with an ask that only the folder makes possible:
Read the last six months of meeting notes in [client folder] and pull
every story they've told about why they started the business.
Specific answer back? It works. Generic answer? The connection or the structure is broken. Fix it now, while the stakes are low.
Step 6: Feed it, continuously
Meeting notes the second a call ends. Brain dumps exactly as messy as they arrive. Decisions, ideas, drafts. A second brain is only ever as good as what you put in it, and the worst setup is the one you stop feeding.
The test your system should pass
Ask this of whatever you build: can it handle one client as easily as it handles five? As easily as fifty? Mine can, because every folder stands alone and they all point back to the same shared rules. If your structure only works because you personally remember where things are, it isn't a system yet.
FAQ
What is an AI second brain? One organized, AI-readable home for everything you know: meeting notes, strategy docs, drafts, decisions, and the rules your AI works by. The difference from a classic second brain is the reader: it's built so your AI tools can pull from it, instead of existing only for future-you.
What's the best app to build a second brain in? The one you'll maintain. GitHub for maximum portability (my pick), Obsidian for a fast solo setup, Notion for team collaboration with real lock-in, Google Drive if you want zero new tools.
Notion or Obsidian for a second brain? Obsidian if it's just you and you want plain files you own. Notion if your team collaborates in it daily and you accept that your content lives in Notion's format.
How long does the setup take? A weekend for the structure, READMEs, and connection. Then it's a feeding habit, not a project: notes and ideas go in continuously or the brain goes stale.