What to Put in Each Brain

The complete decision guide for which knowledge goes where: User Brain, Company Cortex, Customer Brain, Agent Brain, or Space Context.

The Quick Answer

The Decision Flowchart

When you have a piece of knowledge to upload, ask:

1. Is this about a real customer or customers? → Customer Brain
2. Is this about how the org behaves as a whole? → Company Cortex
3. Is this about you as an individual? → User Brain
4. Is this expert content that makes one role world-class? → Agent Brain for that role
5. Is this only needed for current work? → Space Context

User Brain: You as an Individual

Knowledge about you personally. How you think, how you communicate, what you know. This is your identity layer. If an agent working anywhere should know it about you, it goes here.

Put in the User Brain:

• Your personal communication style and tone preferences
• Your origin story and background
• Your individual decisions and the reasons behind them
• Meeting recordings from your strategy calls (auto-ingested via Fathom)
• Personal notes, frameworks, things you want to remember
• "I never use emojis in professional copy" type preferences

Do not put here: How the org operates (Company Cortex). Real customer signals (Customer Brain). Expert content for one role (Agent Brain). Active project details (Space Context).

Company Cortex: How the Org Behaves

Company Cortex is not a wiki. It is the operating mind of the organization: what the company believes, what good means, how agents should collaborate, what to repeat, and what to avoid.

Put in Company Cortex:

• Operating beliefs and company standards
• Collaboration protocols for humans and agents
• Anti-patterns agents should not repeat
• Reusable moves that worked across real work
• Decisions and why direction changed
• Tensions like "move fast" vs. "ask before changing workspace state"
• Retrieval rules that tell agents when a company truth matters

Do not put here: Your personal voice (User Brain). Customer-specific signals (Customer Brain). Expert frameworks for one role (Agent Brain). Active docs/tasks/drafts (Space Context).

Customer Brain: Real Customer Evidence

Knowledge built from actual customer interactions, not invented personas.

Put in the Customer Brain:

• Tagged Fathom recordings of sales calls, customer interviews, support calls
• Notes about a specific customer's preferences, objections, goals
• Survey responses and open-ended feedback
• Customer support thread highlights
• Behavioral signals (what they did, what they bought, what they churned from)
• Untagged general customer intelligence ("our customers care about X")

Do not put here: Top-down marketing avatars you invented (those are active Space/artifact context). Your org's operating standards (Company Cortex).

Agent Brain: Specialist Training

Knowledge that makes one specific agent exceptional at their role. Think "what would I give this person on their first day to make them the best in the world at their job?"

Put in the Agent Brain:

• Expert content in their field (YouTube videos of top performers, books, courses)
• Best-in-class examples of the work they will produce
• Industry frameworks and methodologies for their specialty
• Case studies and benchmarks for their domain
• Platform-specific best practices (if they work on specific channels)

Do not put here: Your business info (User or Company Cortex). Customer evidence (Customer Brain). Active project details (Space Context).

Examples by role:

Space Context: Active Work

Active project knowledge belongs in the Space where work happens. Space Context is what agents use while doing the current work, but it should not all become durable memory.

Keep in Space Context:

• The offer you created for this campaign
• Customer avatar specific to this initiative (declared, top-down)
• Brand theme and visual direction
• Documents, SOPs, and checklists for this project
• Reference content related to this campaign
• Meeting recordings about this project
• Past deliverables worth building on
• Competitor research specific to this market

Promote out only when durable: If a Space teaches an org-wide operating lesson, approve it into Company Cortex. If it teaches something about real customers, promote it into Customer Brain. If it trains one role, move it into Agent Brain.

The Compound Effect

Here is what happens when all five brains are working together:

An agent starts a task for your product launch campaign. They reach into:

1. User Brain "Sefy writes in short sentences, never uses buzzwords, and always leads with data."
2. Company Cortex "For workspace mutations, agents ask before creating tasks. Customer-facing copy should be direct and non-pushy."
3. Customer Brain "Most prospects we have talked to in the last 90 days mention budget pressure as the top concern. Three customers have explicitly asked for an enterprise plan."
4. Agent Brain "I have studied the top researchers in this space. I know proven analysis frameworks and presentation techniques."
5. Space Context "This specific workstream is the Q3 launch. Here is the offer, competitive analysis, pricing strategy, and recent deliverables."

The result is work that sounds like it came from an expert who deeply knows your business, your customers, and this specific project. Not generic output. Informed, contextual, on-brand work.

Moving Knowledge Between Brains

As your projects evolve, you will discover knowledge that belongs in a different brain:

• A Space document turns out to describe how the whole company works → approve it into Company Cortex
• Expert content in the User Brain should really only train one specific agent → move it to their Agent Brain
• An agent produces a framework that is useful for one role → copy it to that Agent Brain
• A note tagged to one contact turns out to apply to a whole customer segment → copy the memory across to a Customer Brain synthesis

Use the copy and move actions from any memory's detail view in the Brain visualization to transfer knowledge where it belongs.