Walks

See what an agent actually did.

Test With AI Agents

Optional. Run a live model or a scan simulation against a task. This does not change your audit scores.

Customize Audit — optional channels, agent walks, or extra reports.

What is this audit?

Test With AI Agents

Optional walks against a task. This is a testing feature, not a fifth optimization service.

Optional walks against a task. Outcomes are Live test, Simulation, Unsupported, or Inconclusive.

Why this audit matters

A score does not show what an agent actually did. A walk shows the path, the answer, and the gap — without claiming a consumer app ran unless it did.

Who it is for

Teams that want to see a task completed or fail, then copy a fix prompt.

How SiteGlue evaluates the website

Claude and Grok can run as models on SiteGlue when a key is present. Other named agents use a scan simulation. That is not the consumer app. Walks do not change ARO, CRO, AEO, or SEO scores.

The formula behind the score

Walks are not scored. They never change ARO, CRO, AEO, or SEO.

Recommended improvements

  • Fix the pages the walk could not complete
  • Do not treat a simulation as a Claude or Muse app test
  • Use Improve Your Score for the underlying audit gaps
  • Live test: a model on SiteGlue fetched public pages
  • Simulation: SiteGlue replayed the scan, not the consumer app
  • Unsupported or Inconclusive when there is no runtime or not enough pages

Frequently asked questions

Did Claude test my site?

Only if the result is labeled Live test and the agent was Claude. Otherwise it is a simulation or unsupported.

Does a walk change my scores?

No. Walks are a separate capability. ARO, CRO, and AEO stay independent.

Customize Audit — optional channels, agent walks, or extra reports.