Receipts give AI work a portable evidence link.
Use your existing observability system for health, timing, alerting, and debugging. Add an execution receipt when one AI tool call needs a stable record that a teammate, client, or another agent can inspect.
How the pieces fit
Show the path and timing of a request across services, spans, and dependencies.
Capture structured or textual events that help an operator investigate behavior and errors.
Track aggregate measurements such as latency, error rate, throughput, and saturation.
Package one AI tool call into a stable, inspectable record with request context, result, timestamp, and hash commitment.
Receipts are not a tracing replacement
A trace answers how a request moved through a system. A log answers what an operator recorded. A metric answers how a system behaved over time. A receipt answers what one AI execution returned and gives another person or agent a focused record to inspect.
A practical review sequence
- Find the trace or request identifier in your normal observability system.
- Open the matching receipt and compare the requested operation with the observed result.
- Check the timestamp, tool arguments, evidence, and SHA-256 commitment.
- Report only what the evidence supports; a receipt does not prove an external outcome the system did not observe.