Read your report
Your report has two halves: the Agent-Readiness Score, which is the state we help you control, and the four-engine citation check, which is the proof. Here is how to read each part.
Readiness is what you act on; citation is what you watch. The score and its gaps tell you what to fix. The engine check shows what the engines did with your page today. Keep the two distinct and the report reads cleanly.
The Agent-Readiness Score
A single number out of 100 for how readable and parseable the page is to AI engines. It rolls up a set of readiness dimensions, each marked pass, partial, absent, or n/a. The dimensions are the actionable part: each one maps to a specific gap you can close.
The two pillars
- Pillar A, your site as a source. Can the engines read you and quote you? Schema, clean entities, answerable content, an llms file. This is the headline today, and what the score leads with.
- Pillar B, your site as a participant. Can agents act with you, not just read you? Machine-readable pricing and policy, agent manifests. This is the next layer; the model already scores it so your report does not need rewriting as agentic commerce matures.
The gaps
Below the score is the ranked list of what pulled it down: the missing or partial structure on the page. On WordPress, most of these become fixes you can approve and apply. See Approve fixes.
The four-engine citation check
For each of ChatGPT, Perplexity, Gemini, and Claude, the report shows whether the engine returned your page for a relevant question, and how many sources it pulled. This is the proof layer: it is a measurement of what happened, not a promise of what will happen. Engines are nondeterministic, so this is a reading you watch over time, not a guarantee.
Who else is cited
The report also surfaces the other pages an engine returned alongside or instead of yours. It is a way to make the gap concrete: if an engine returned a comparable page and not yours, the structure that page has is often exactly what your readiness score flags as missing. We surface this as a diagnostic, not as something we monitor for you.
On a paid report: the sim panel and ranked fixes
The paid report adds a 20-agent simulation panel and a confidence reading. The panel predicts how engines should rank your page; the confidence (shown as high, moderate, or weak) is how closely those predictions line up with the real engine results. It also ranks each recommendation by its predicted readiness lift, shows a per-URL audit, and exposes the raw per-engine evidence so you can check the reading yourself.
Ready to close the gaps? Approve fixes walks through applying the agent-ready layer to your site.