Agent trace
A record of one agent run, made of spans for model calls, tool calls, handoffs and other steps.
An agent trace records one end-to-end run of an agent as a set of spans, each covering an operation with a start and end time, such as a model call, a tool call, a handoff or a guardrail check. OpenTelemetry's generative AI conventions, still in development, define spans for invoking agents and workflows, planning and executing tools, and make capture of message content opt-in. Putting the trace identifier on the resulting payment and case record lets one search show what happened.
Agent Minute explains this term on 4 March 2027.
Related terms
Audit trailA chronological record of who accessed a system and what operations they performed, enough to reconstruct events.Automatic loggingBuilt-in recording of events while a high-risk AI system runs, so its operation can be traced afterwards.Tamper-evident logA log built so that any later change, deletion or reordering of entries can be detected.Evidence packA bundle of records that reconstructs one agent action: identity, mandate, authentication, inputs, calls, outputs and approvals.