ADO (Agent Discovery Optimization)
Agent Discovery Optimization (ADO) is the emerging discipline of making an API discoverable and legible to AI agents: publishing machine-readable entry points such as llms.txt, AGENTS.md, an OpenAPI spec, and MCP surfaces, and serving documentation that renders on a plain HTTP fetch. It applies the optimization instinct behind AEO and GEO to the surface agents use when choosing what to build on — the API.
APIs earn their own discipline because their agent readers differ from an answer engine's. The agents that matter here are the ones scaffolding software — coding assistants and integration builders that must locate the reference, parse the contract, and construct working requests. They arrive over plain HTTP, skip JavaScript, and navigate by declared entry points, so the work is specific: route them to the reference, give them a parseable spec, brief them explicitly.
ADO also inherits a verification problem. Teams ship an llms.txt or an AGENTS.md and have no way to confirm it changed anything an agent experiences — the work can feel like cargo-culting. A discipline needs a measurement: a way to check that the shipped artifact actually made the API easier for an agent to find and operate.
This is Discry's home ground. The Discry Score measures the outcome ADO work aims at — discovery, meaning a machine-legible surface is present and correctly routed on a plain fetch, and comprehension, meaning real models can operate the API from the docs, graded mechanically. That gives ADO its definition of done and its verification loop: ship the artifact, and the measurement shows whether agents can now find and understand what you published.
What does an AI agent make of your API?
Find out in about a minute — no signup.
Discry your API — free