Topics · Evidence gap
A later reviewer can check what action was signed, under which authority, and at what time without trusting the agent dashboard narrative.
Concrete scenario
An underwriting agent approves 400 policy changes overnight using tool calls across CRM, pricing, and document services. A dispute targets one approval made at 02:14. The operator log shows a green checkmark, but the exported audit trail omits the policy version hash and the delegated authority scope active at execution time.
Problem
Agents act faster than human audit teams can follow. Dashboards summarize outcomes, but disputes need the signed action, authority, and policy context that existed at execution time — not a post-hoc summary.
Mechanism
The accountable action is minted as Verifiable Machine Activity with policy-bound claim fields and routing profiles at execution time. The receipt preserves issuer key, timestamp, and structured context as signed bytes rather than as a narrated log line inside one SaaS product.
Verifiable outcome
Scope boundary
A receipt attests to what was signed at execution — not whether the policy was wise, whether every tool call in a chain was captured, or whether the agent behaved safely in environments without receipt instrumentation.