Decision Provenance Is Not Authority Provenance.
You can trace an AI outcome all the way back through its data, model, policy, and controls — and still not know whether the actor was entitled to cause the effect.
Explaining how an outcome was produced is not proving anyone was allowed to act on it.
What decision provenance gives you
Decision provenance answers "how was this outcome produced?" — and modern tooling answers it well.
- A lineage from source data through the model to the final output.
- The policy and control checks the decision passed on its way.
- An explanation of the reasoning, with the inputs that drove it.
This is the production trail: data → model → policy → control → outcome. It is exactly what a team should keep, and it makes the output inspectable. But it describes the outcome; it never names the mandate behind the act.
Where it stops
Decision provenance traces the outcome; authority provenance traces the entitlement.
The enterprise research landscape converges on four properties that make an AI output trustworthy — accuracy, provenance, explainability, and self-correction. All four are necessary, and none of them is authority. A recommendation can be 95% accurate, fully provenanced, cleanly explained, and self-correcting, and still be one that this actor had no right to turn into an action — for the wrong principal, outside its purpose, past its expiry. A trustworthy output is not an authorised act, and the two are proven by different chains.
- It records how the outcome was made; it never records who was entitled to cause it.
- It follows data → model → policy → control; it does not follow institution → mandate → office → principal → delegation → agent → action.
- A perfectly explained decision, acted on without authority, is still a breach.
The line
Authority provenance proves why the actor was institutionally entitled to cause the effect.
KYE Protocol™ resolves the second chain — the one that starts at an institution and descends through mandate, office, principal, and delegation to the agent taking the act — and checks the request against the delegated authority the principal actually holds, within purpose and scope, at the boundary. It returns admit or deny the instant before the action and seals that verdict into a replay-verifiable Evidence Pack™. That is Authority Finality™: decision provenance explains the outcome; authority provenance proves the right to bring it about.
Honest scope
KYE Protocol™ proves authority, not the quality of your output.
Whether a decision was accurate, well-provenanced, and cleanly explained is the work of your model, your lineage, and your evaluation stack — and all of it stays essential. KYE Protocol™ consumes that trustworthy output and closes the gap it cannot close for itself: the actor's institutional entitlement to act, bound to the principal and provable to an auditor from public keys alone. Keep the decision trail. Add the authority chain.
Explaining how an outcome was produced is not proving anyone was allowed to act on it.