The trust layer for consequential action.
In an age of agents, legitimacy becomes infrastructure. Every consequential action — a payment released, an account opened, an agent acting for you, a claim published — needs a layer that decides whether it may proceed, records why, and makes that answer final and contestable. That layer is missing from the stack. KYE Protocol™ is that layer. You adopt AI at full speed; you keep legitimate control of what it is allowed to do.
Why a trust layer, and why now
You are being asked to let software act on your behalf — to move money, sign a customer, answer a regulator, publish a position. When it acts, three questions land on you, the CEO and the board: who authorised that action, within what purpose and scope, and can you prove it after the fact? Identity and access management answers who logged in. Model governance answers was the model well-behaved. Neither answers was the action legitimate. The trust layer does.
A trust layer earns its name by being provable to a sceptic. Each decision KYE Protocol™ makes is sealed, signed, and replay-derivable from public keys alone — so a third party validates it against 170 declared frameworks (NIST 800-207, ISO/IEC 42001, the EU AI Act™ Article 14) without trusting the vendor:
- Decide — admit or refuse the action at the moment it is attempted, against the purpose it was authorised for.
- Record — capture why, as a signed Evidence Pack™ that survives the agent, the model, and the vendor.
- Make final — resolve who held authority to a sealed answer that a regulator, an auditor, or a counterparty can replay.
The chain of legitimacy
A consequential action does not become legitimate in one step. It travels a chain, and the trust layer governs every link — so authority never leaks between them. Read top to bottom; each link is the input to the next, and the whole resolves to finality:
- Purpose — the declared reason the action is allowed to exist, scoped and time-bound. KYE Protocol™ calls this Purpose Permission™.
- Authority — the delegated permission that licenses the act, traced across every principal in the chain.
- Delegation — how that authority passes to an agent, a tool, or a teammate without widening on the way.
- Execution — the action itself, admitted or refused at the moment it is attempted, not reviewed weeks later.
- Evidence — the signed record of what was decided and why, contestable by the person it affects.
- Finality — a sealed answer that replays from public keys alone. KYE Protocol™ calls this Authority Finality™.
One doctrine, many implementations
The trust layer is the umbrella; the rest of KYE Protocol™ is how it is delivered. You do not buy a pile of features — you adopt one layer, and everything else is an implementation of it. Each surface below is a projection of the same doctrine, never a separate product to integrate by hand:
- Agentic Governance™ — the category the doctrine defines: governance for humans, AI agents, and hybrid authority.
- Evidence Pack™ — the signed artefact that makes any decision replay-provable to a regulator or counterparty.
- KYE GovernedUI™ — the human-control surface where a CISO or counsel approves, holds, or refuses an agent's authority.
- Customer sovereignty — you hold the keys; the proof is yours, portable, and does not depend on KYE Protocol™ staying in the room.
Adopt the layer, keep control.
If your team is putting agents into a consequential path — payments, onboarding, advice, anything a regulator can ask about — the trust layer is what lets you say yes without losing legitimate control. Start with one governed pilot on a single real action, and leave with an Evidence Pack™ you can hand to your board, your auditor, or your regulator.