Open-source playbook · 01 of 20 · Start here
Your company’s knowledge lives in a dozen tools and a few people’s heads, so AI gives generic answers.
The systemOne organized, always-current memory of your customers, decisions, processes and voice that every person and every AI system can ask.
// The problem
The answer exists. It’s just spread across tools, threads and people. So someone searches for half an hour, asks two colleagues and still gets half of it. AI tools have the same problem: without your context, they give generic answers.
TodaySearch four tools, ask two people, hope it’s current.
With a Company BrainOne answer in seconds, with the source it came from.
// How it works
Nothing moves out of your tools. The brain reads them, keeps a clean, linked copy of what matters, and answers from it.
1What you already have
Synced automatically, every hour
2The brain: six drawers, every page linked to the pages it relates to
Customers
People & owners
Decisions
Processes
Products & pricing
Voice & brand
Read by people and by AI
3People ask in Slack or Teams
“@brain what’s our refund policy for annual plans?”
3AI systems read the same pages
// What it feels like
Sample questions and answers (fictional company). Every answer links back to where it came from, so people trust it and check it in one click.
A named onboarding lead and a kickoff within 10 working days of signature. Priya confirmed it on the 14 March call, and it’s in section 4 of the signed proposal.
Acme · call notes, 14 MarAcme proposal v3, §4
Up to 10% without approval. Anything above needs the CFO, per the pricing decision from 2 September.
Decision log · 2 SepPricing policy
Daniel. He has run 7 of the 9 healthcare accounts and wrote the HIPAA checklist.
Accounts by ownerHIPAA checklist
// The build
Week 1
List the questions people ask most and the tools the answers live in. Choose the first six drawers.
Week 2
Sync the sources. Turn raw files into short, linked pages: one per customer, decision, process and product.
Week 3
Let people ask in Slack or Teams, and let your AI systems read the same pages, with sources shown.
Week 4
Give every drawer an owner, refresh it automatically and flag pages that go stale.
// For your tech lead
Start small: one team’s drawers on a no-code setup. Move to a production build once people use it every day.
| Job | Use | Or |
|---|---|---|
| Where the knowledge lives today | Google Drive + Slack | SharePoint + Teams |
| Keeping the brain in sync | n8n | Make, Zapier, or a nightly script |
| The brain itself | Linked markdown pages in a repo or Notion | Confluence, a vector database |
| Answering questions | Claude (Projects or API) | ChatGPT, Gemini |
| Asking from chat | A Slack app | A Teams bot |
Start simple2–3 days
Upload your key docs into a Claude or ChatGPT project for one team. No code, and you learn which questions people actually ask.
Production2–4 weeks
n8n syncs your tools into linked pages every hour; a Slack app and your AI systems read them through one connector.
Custom4–8 weeks
An engineer adds search across everything with permission checks per person, plus usage analytics.
An ops or RevOps owner who knows where things live, and an engineer for the connectors in weeks 1–2. After that, each drawer has an owner who spends about 30 minutes a week on it.
Needs
Nothing. Start here.
Next to build
All 19 other systems read from it.
// Want help building it?
In 30 minutes with our founder, we’ll look at your tools and your week, and work out what your first AI systems should be and how long they’ll take.
Book a Strategy Call
Free 30-minute call with our founder · No commitmentPallav Kaushish Co-Founder, AIxGrowthPick a time that works
No times are open right now. Please check back soon.
You're booked
While you wait, watch the 13-minute demo