ChainIT Releases Runtime Compliance Protocol for AI Agent Transactions

ChainIT announced in a press release a protocol that checks whether current compliance evidence and organizational policy permit a specific transaction. The protocol covers actions initiated by AI agents, authorized people, organizations and controlled workflows.

The system records the source, provenance, observation time, confidence, permitted purpose and privacy classification of evidence. A versioned policy profile selects applicable requirements, while rules and workflow engines produce a signed decision with outcomes such as allow, review, hold, reject or prohibit.

The protocol can reassess compliance while a transaction is in progress and maintain an append only evidence record. A compliance approval does not provide payment authority on its own. Identity, authority, scope, approvals, capacity, transaction details and a credential for one use must also be verified before funds move.