Trust but Verify
SanctifAI Trust is the audit trail for human oversight of agentic AI: every review, approval, and escalation captured as verifiable, tamper-evident Proof of Human participation.
Loading latest certificate...
How it works
SanctifAI Trust adds human accountability to agentic processes.
Trust sits between agentic work and accountable approval. It creates a durable record that can follow the work into audits, customer portals, internal dashboards, and third-party workflows.
An agent requests a human.
When an agentic workflow crosses a policy threshold, SanctifAI issues a cryptographic challenge — task ID, input/output hashes, and the control that requires a human. It surfaces to a qualified reviewer embedded in your app or through the Chrome extension.
Anatomy of a proof
What's actually in a Proof-of-Human certificate.
Every field is here because an auditor, a regulator, or opposing counsel will eventually ask for it. Nothing extra. No PII.
No names, no emails, no document content, no biometrics. Identity is dereferenced under legal hold via a separate governed channel.
Your tenant database plus an Ethereum Attestation Service anchor. SanctifAI cannot edit a sealed attestation. Neither can you.
As long as a single mirror of the chain exists. Verifier keys rotate; old certs remain verifiable against pinned historical keys.
Trust but Verify
In the end, every responsible-AI obligation reduces to one question — was human agency real and present when it mattered? Verifiable attestation is what makes the answer checkable.