“Company brain” is an emerging label for shared organizational context that people and AI can use. It is not a standardized architecture. Its usefulness depends on reliable sources, clear ownership, freshness, and permission-aware retrieval.
A useful idea with a loose name
Sentra describes its company brain as governed organizational context shared by people and agents. That is a vendor’s framing, not a universally accepted definition. The underlying product question is practical: how can a team retrieve the knowledge it needs without rebuilding context every time?
Anthropic’s context-engineering work offers another useful lens: the information presented to an agent is an engineering choice. Connecting more tools does not automatically produce better context.
Define what deserves to become knowledge
A conversation, a decision, and an approved policy are different objects. Preserve those differences. A comment in a meeting should not silently become an authoritative rule.
For a decision, capture its owner, date, rationale, scope, and whether it supersedes something earlier. For a reusable instruction, identify the accountable team and the event that should trigger a review. These are recommended design fields, not mandatory industry standards.
Make permissions travel with the context
A shared index must not become a shortcut around source permissions. Design retrieval and action boundaries with the identity and authorized scope of the person using the system.
Test the experience with contrasting roles and changed access. A useful acceptance scenario asks whether someone who loses access to a source can still retrieve its sensitive content through derived answers. Implementation needs to consider copies, caches, and summaries.
Treat the brain as a maintained product
Start with one decision workflow and measure whether people can find a dependable answer. Assign owners to gaps. Track stale or conflicting answers, and make corrections discoverable.
My KBaP lens puts this maintenance at the center. Knowledge needs users, quality standards, and a feedback loop. A large connected archive is an input; reliable decisions are the outcome.
Questions worth asking
Is a company brain the same as a knowledge base?
The term often includes retrieval, relationships, and shared context around a knowledge base. Definitions vary by vendor; evaluate the concrete capabilities.
What is the smallest useful starting point?
Choose one repeated decision, identify its authoritative sources, and test whether authorized users get a fresh, dependable answer.
Sources & further reading
These sources inform the context. Recommendations and frameworks are editorial analysis, not claims of proven results. AI-assisted draft prepared for Yaron’s review; personal experience claims require his confirmation.