Framework · Maturity
The AI Trust Maturity Model
Seven levels from experimental AI use to autonomous, self-verifying trust infrastructure — with the concrete capability that defines each.
01
Seven levels of maturity
- Level 1 — Experimental: Ad-hoc AI use with no formal governance.
- Level 2 — Managed: Basic controls and human ownership begin to appear.
- Level 3 — Defined: Standardised policies (Standards & SOPs) and processes are documented.
- Level 4 — Governed: Every agent action passes a mandate + Standards + SOP gate — allow / observe / block / escalate — with signed evidence.
- Level 5 — Trust Enabled: Continuous TrustOps: a live trust index, evidence anchoring and human-in-the-loop escalation run across the estate.
- Level 6 — Federated: Verifiable identities and trust records (did:hedera / did:key, HCS-28) are exchanged across organisations and borders.
- Level 7 — Autonomous Trust Infrastructure: Trust is programmable and self-verifying — signed, anchored on-chain, and re-derivable by anyone with no trust in the issuer.
02
How Metamynd moves you up the model
- Governed (4) — turn on the mandate + Standards + SOP gate; every action is allow / observe / block / escalate with signed evidence
- Trust Enabled (5) — run TrustOps: a live HCS-28 Trust Index, evidence anchoring, escalation review
- Federated (6) — exchange did:hedera / did:key identities and HCS-28 trust records across orgs
- Autonomous (7) — publish signed, on-chain trust records anyone can re-derive and verify offline
03
Using the model
Assess your current level, identify the gap to the next, and use TrustOps to progress. Most organisations begin between Managed and Defined and target Trust Enabled — the point at which trust is measured continuously, not asserted once.