The five areas of agent readiness: find, read, act, pay, trust
In short
Agent readiness has five areas. Find asks whether agents can reach you, read whether they can understand you, act whether they can finish a task, pay whether they can complete the purchase and trust whether you can tell who is who. The self-check scores each area and the weakest one caps your level, so the areas are worked in order.
The areas and their questions
| Area | The question | What it covers | Self-check questions |
|---|---|---|---|
| Find | Can agents find you? | Whether AI crawlers and assistants are allowed in, and whether they can locate your important pages. | 1, 2, 3, 4 |
| Read | Can agents read you? | Whether products, prices, policies and articles are available as text an agent can parse. | 5, 6, 7, 8 |
| Act | Can agents act on your site? | Whether the main task (buy, book, subscribe, request a quote) can be completed through standard forms or a documented interface. | 9, 10, 11, 12 |
| Pay | Can agents pay? | Whether full prices are visible before the last step, and whether your payments team has a position on purchases an agent starts. | 13, 14, 15 |
| Trust | Can agents be trusted, and trust you? | Whether you can tell verified agents from unknown automation, and whether your terms and policies are public. | 16, 17, 18 |
The order to work in
Work find and read first. In Ora's experiment the two changes that stopped an agent were both about these areas: an answer hidden in script, and a refusal returned to agent user agents. Act and pay follow, because there is little point improving checkout for a visitor that cannot read the product page. Trust runs through all of them, since it decides who you let in and who answers for the policy.
- 19 Oct 2026
Ora's 19-configuration experiment (190 probes) found only two changes stopped retrieval: an answer hidden in JavaScript (0 of 5 for plain fetch) and a 403 to agent user agents (0 of 5 for both clients).
- 19 Oct 2026
Agents reached docs pages in 83 percent of traced runs and the home page in 69 percent; .well-known files 23 percent, openapi.json 21 percent where shipped, and sitemap.xml 4 percent.
Why the weakest area caps the level
An agent's journey is a chain. A site that is easy to read and impossible to pay is still a site where the job does not finish. The scoring rule reflects that: the final level can be at most one above the weakest area's level. The readiness levels page has the worked example.
Related
- The self-checkEighteen questions across the five areas.
- Readiness levelsHow the five areas become a level.
Sources
- AgentReady.org guide (agentready.org), read 4 Nov 2026. Vendor claim.
Last reviewed