OpenTelemetry GenAI Semantic Conventions (agents and MCP) vs Agent Audit Trail (AAT)
The map lists them as complementary: sites and agents commonly implement both.
| OpenTelemetry GenAI Semantic Conventions (agents and MCP) | Agent Audit Trail AAT | |
|---|---|---|
| Lifecycle step | 9 Account: Prove what happened; be judged | 9 Account: Prove what happened; be judged |
| Answers | What span, metric and attribute names should agent frameworks and MCP clients and servers emit, so traces of agent runs and tool calls line up across vendors? | Is there a standard, tamper-evident log format any autonomous AI agent can write its actions to, so a regulator, auditor or victim can later reconstruct exactly what it did and why? |
| Governed by | OpenTelemetry project (backed by the CNCF); developed in the open-telemetry/semantic-conventions-genai repository, which extends the core OpenTelemetry Semantic Conventions | IETF individual Internet-Draft; not adopted by any IETF working group, 'not endorsed by the IETF' and has 'no formal standing' in the IETF process (datatracker's own disclaimer) |
| Status (issuer's words) | Status: Development (OpenTelemetry document status) on both the agent-spans and MCP documents; the repository has no tagged release | Active Internet-Draft (individual submission), Intended status: Standards Track; not a Working Group document |
| Version | Development (unreleased; repository has no tags); builds on core semantic conventions v1.44.0 (2026-10-19) | draft-sharif-agent-audit-trail-08 (2026-10-13) |
| Form / transport | document format | document format |
| Found at | None at the web layer; the conventions name attributes on OpenTelemetry spans and metrics that an agent's own instrumentation exports to its operator's tracing backend | none |
| Implementation | prototype | none found |
| Evidence | 3 verified | 3 verified |
| Primary use case | Audit what an agent did | Audit what an agent did |
| Last verified | 2026-10-23 (current) | 2026-10-23 (current) |
| Specification | https://github.com/open-telemetry/semantic-conventions-genai/blob/main/docs/gen-ai/gen-ai-agent-spans.md | https://datatracker.ietf.org/doc/html/draft-sharif-agent-audit-trail-08 |
When to use which
- Choose OpenTelemetry GenAI Semantic Conventions (agents and MCP) when your question is “What span, metric and attribute names should agent frameworks and MCP clients and servers emit, so traces of agent runs and tool calls line up across vendors?” and you accept its maturity: status "Status: Development (OpenTelemetry document status) on both the agent-spans and MCP documents; the repository has no tagged release", implementation prototype.
- Choose Agent Audit Trail when your question is “Is there a standard, tamper-evident log format any autonomous AI agent can write its actions to, so a regulator, auditor or victim can later reconstruct exactly what it did and why?” and you accept its maturity: status "Active Internet-Draft (individual submission), Intended status: Standards Track; not a Working Group document", implementation none found.
Read the full articles
- OpenTelemetry GenAI Semantic Conventions (agents and MCP): What span, metric and attribute names should agent frameworks and MCP clients and servers emit, so traces of agent runs and tool calls line up across vendors?
- Agent Audit Trail: Is there a standard, tamper-evident log format any autonomous AI agent can write its actions to, so a regulator, auditor or victim can later reconstruct exactly what it did and why?