What is in an AI governance record?

The six questions the record is built to answer

Start with what the record is for. It is not a pile of logs — it is organized around the questions a council, an auditor, or a regulator actually asks:

An AI governance record is built to answer six questions: what AI agents existed, what each could do, what each actually did, in what context, what the outcome was, and which policies or controls applied.

Everything in the record earns its place by helping answer one of those six. That is why it is a record, not just storage.

Two lanes: what we witness, what you file

The record is fed from two directions, and keeping them distinct is what makes it credible.

• What we witness — read-only. Agentica continuously maps every AI agent in your business environment and records what it did and when. By default: “we can see that an AI tool modified a file at 2:00 PM and who owns it — never what is inside. You decide, data class by data class, where the record goes deeper.” Metadata and signals, witnessed, never acted on.
• What you file — your governance acts. Your policies, assessments, approvals, and incident entries are filed on Agentica’s templates and entered into the record as dated exhibits — sitting in the same tamper-proof, independently held history as the machine signals. The immutable history carries both, side by side.

The first lane is the evidence of what your AI did; the second is the evidence of how you governed it. A reasonable-diligence question needs both.

What governance instruments are

The templates behind that second lane are governance instruments: a versioned, framework-mapped library of best practices, checklists, procedures, and templates — generic, self-serve, authored once, and adapted and adopted by you. Adopting one is itself a record event: the adopted artifact is sealed into the record, its review dates tracked, its updates versioned.

One honest boundary: sealing is not verifying. We record what you filed and when — we never attest that its content is correct or complete. Your officers remain accountable for the governance itself. Instruments are generic best practice, never a customer-specific prescription, and they ship only inside the record, never as a standalone product.

Agentica, an AI Governance company in Montréal, is the auditable system of record for your AI workforce: a register to improve your posture, and the record to prove your diligence.

Agentica provides AI-governance evidence and risk intelligence. It does not constitute legal advice or compliance certification.