Cookies, Syslog & T&Cs vs Agent Attribution, OTel GenAI & SCITT
HTTP referrer cookies, server access logs & human Terms of Service is the classic web primitive built for human browsers and pre-integrated services (Track human affiliate referrals in browsers and log HTTP hits for debugging), whereas Agent Attribution Protocol (AAP) and its step 9 (Account) peers provide the machine-native agentic surface (none) so autonomous agents can answer: "When an AI agent recommends a product and a sale happens later, on another device or session, can anyone prove which agent (and which recommendation) actually drove it?" Across the agenticweb.wiki conformance panel, Agent Attribution Protocol is shipping (multi-implementation in production).
How Cookies, Syslog & T&Cs compares to Agent Attribution, OTel GenAI & SCITT at step 9 (Account): why the classic primitive breaks for software agents, and how sites deploy both together.
| Dimension | HTTP referrer cookies, server access logs & human Terms of Service Classic web | Agent Attribution Protocol AAP Agentic web |
|---|---|---|
| Primary consumer | Human driving a web browser or pre-integrated developer code | Autonomous software agent, coding harness, or AI search/RAG crawler |
| Lifecycle step | Classic web baseline for step 9 (Account) | 9 Account: Prove what happened; be judged |
| Core question answered | Track human affiliate referrals in browsers and log HTTP hits for debugging. | When an AI agent recommends a product and a sale happens later, on another device or session, can anyone prove which agent (and which recommendation) actually drove it? |
| Wire format & entrypoint | Set-Cookie / Referer headers + unstructured access.log lines | endpoint at none |
| Failure mode for agents | Agents do not carry third-party affiliate cookies across chat summaries, standard web logs cannot prove which prompt or tool call triggered an action, and an autonomous software process is not a legal person capable of clicking 'I Agree'. | Solved natively: The Agent Attribution Protocol (AAP) is a commerce-attribution scheme built by Rako, a UK company, so that when an AI shopping agent recommends a product and a user buys it later -- on a different device, browser, or after a delay -- the agent's role in driving that sale can be proven and a commission paid. |
| Governance & status | Established IETF / W3C / industry baseline | Single vendor (Rako); the spec text is published under Apache-2.0, but issuance, verification and settlement are centrally operated by Rako, which the spec calls the 'Protocol Authority' — "The open standard for AI agent affiliate attribution" (vendor's own description); v1.0 spec published, centrally operated |
| Live adoption (/check/ census) | Universal across the classic web | shipping — multi-implementation in production |
| Primary specification | https://www.rfc-editor.org/rfc/rfc6265 | https://github.com/rakohq/aap-spec |
Why Cookies, Syslog & T&Cs breaks when the caller is an agent
Agents do not carry third-party affiliate cookies across chat summaries, standard web logs cannot prove which prompt or tool call triggered an action, and an autonomous software process is not a legal person capable of clicking 'I Agree'.
When to use which (and how they work together)
- Keep HTTP referrer cookies, server access logs & human Terms of Service: You are measuring traditional browser session analytics and human affiliate click-throughs.
- Add Agent Attribution Protocol (AAP): You need cryptographic publisher attribution when an agent cites or transacts on your content (Agent Attribution Protocol), standardized GenAI telemetry spans (OTel GenAI), or tamper-evident transparency ledgers (SCITT / Agent Audit Trail).
All agentic web standards replacing or extending Cookies, Syslog & T&Cs
- Agent Attribution Protocol (AAP) — shipping: When an AI agent recommends a product and a sale happens later, on another device or session, can anyone prove which agent (and which recommendation) actually drove it?
- OpenTelemetry GenAI Semantic Conventions (agents and MCP) — prototype: 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 (AAT) — none found: 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?
- AI Agent Execution Profile of SCITT (AIR) — none found: How can an agent operator produce independently verifiable, tamper-evident records of an AI agent's material actions using SCITT signed statements and an independent transparency service?
- An Audit Record Format for AI Agent Authorization Decisions (Agent Audit Record) — none found: What should a tamper-evident record of an agent's authorization decision contain?
- Agent Decision Commitments for Payments (ADC) — none found: How can an agent's payment decision, including the content it gave the model, be committed to before payment and later disclosed to dispute participants for verification?