LANBrain
Back to homeDocumentation

Guide

Learning paths

There is no single right way to use your brain. There are three common paths, and they map to the kind of person you are right now. Pick the one that fits, follow it until it stops being useful, then look at the next path.

You can switch paths any time. The brain does not change. Only your relationship with it does.

Path 1: the curious beginner

You started your brain recently. You are not sure what to write first. You want something to feel useful by the end of today.

Five small steps, in order:

  1. Write your about-me doc. Open about-me/about-me.md. Spend twenty minutes telling the brain who you are. What you do for work, what you study, what tools you use, the problems you want help with. Write like you are talking to a new colleague. The template inside the file walks you through it.
  2. Capture five sources. Find five things you already have lying around: your CV, a recent important email, a Notion page you reference often, a PDF of a contract, a screenshot of something useful. Drop them into raw/.
  3. Run the first ingest. Ask your AI chat to ingest the five sources and write the first wiki pages.
  4. Ask the brain three questions. Open your chat with the brain connected. Ask "what do you know about me?" Then "what are my main projects?" Then anything you actually want to know. The answers will be thin (you only gave it five sources), but you will feel the brain working.
  5. Add five more sources. Each week. Quietly. The brain compounds. By week eight you will have a thinking partner.

That is the whole beginner arc. The brain becomes useful before it becomes deep. Do not skip step 1. The about-me doc is the highest-leverage page you will ever write.

Path 2: the power user

You have been using the brain for a few weeks. The basics feel comfortable. You want to make it do more for you.

Five places to go deeper:

  1. Build a skill. A skill is a small markdown file that teaches the AI to do something specific for you. "Summarise any meeting transcript into the format I prefer." "Extract action items from my Slack export." "Draft a thank-you email after every coffee chat." Ask your AI chat for examples and copy one, edit it, save it. You just gave the brain a new tool.
  2. Connect more MCPs. Your AI chat has native integrations (Gmail, Calendar, Notion, Drive, Slack, GitHub, and more depending on the provider). Each integration becomes an input or an output for your brain. Start with one. Use it for a week. Add the next.
  3. Run automations. Move from "I ask the brain to do X" to "the brain does X on a schedule". Daily summaries. Weekly reports. Auto-drafts of emails you write repeatedly. Your AI chat probably has a scheduling feature.
  4. Customise CLAUDE.md. The file is the constitution of your brain. Out of the box it is generic. After a month you will know what tweaks you want: a custom voice rule, a specific format for source summaries, a folder convention that suits your work. Edit the file. Every AI conversation after that update behaves differently.
  5. Read what other people are doing. The LAN docs at https://www.lanbrain.ai/docs publish patterns from other users (sanitised). Worth a monthly skim. You will steal good ideas.

The power user phase is where the brain stops being a tool and starts being a system. Take your time. Skills compound.

Path 3: the team operator

You are running a team or a company. You want the same brain methodology to work across multiple people, with role-based access and shared context.

This path is what the corporate brain product is built for. The personal brain is the starting point. The corporate features layer on top.

Four bullets to point you in the right direction:

  1. Multi-vault setup. Each person on the team gets their own personal brain. A separate shared vault holds the team's collective knowledge: decisions, processes, customer data, deals. The personal brains link into the shared vault. The shared vault never sees the personal brains.
  2. Role policies. A simple markdown file inside each shared vault (policies/access-control.md) defines who can read what. Sales reads wiki/customers/. Engineering reads wiki/architecture/. Everyone reads wiki/people/. The LAN MCP server enforces these policies whenever an AI client tries to read.
  3. Corporate ingestion. For most teams, "done-for-you" ingestion is the right starting point. You hand LAN your data (Slack export, Notion workspace, Drive folder), LAN ingests it into your shared vault, your team queries it from their AI tools. After a few months, when the team is comfortable, you can switch to self-serve ingestion.
  4. Operator handbook. The corporate playbook covers governance, retention, onboarding new employees into the brain, offboarding, and audit. The full handbook lives in the docs.

If this is your path, get in touch with the LAN team from your dashboard. The corporate brain has a different setup flow than the personal one, and a coach walks you through it.

Where to go for more

The methodology is alive. The docs online stay current as LAN ships updates.