Skip to content
Axon · Governance

Hold the brand truth every module reads.

The system of record for what is true about the brand. Sense grades against it. Studio cites from it. Nothing enters Axon lights-out.

01 / How it works

Cited in. Confirmed by a human.

Guidelines, decks, and legal documents come in with every fact tied to the page it came from. Nothing is paraphrased into truth. Every register is confirmed by a human before the loop can act on it. Confirm, don't compose.

02 / The registers

What the brand confirms, the models are taught.

IdentityNames, marks, handles, and the language the brand owns.
IntentThe position to defend in the answer.
GuardrailsClaims never made, and claims that require legal sign-off.
Approved factsStatements the models may be taught, each cited to its source page.
03 / Source of truth

Read by both. Written by neither.

Brand + LegalConfirm & write. The only pens in the room.
SenseReads to grade, so “correct” means something.
StudioReads to cite, so it can only say what's confirmed.
The audit trail always leads back to a human who confirmed the fact
The loop

Written to forward to legal.

The security documentation pack, tenancy, data handling, and the human gate as a control, is available on request.