Home / How it works

How it works

From a free check to verified fixes.

Find, Understand and Use are three different questions. We answer each with its own evidence, fix what is broken, and check again so you can see what changed.

The process

Measure → Diagnose → Fix → Verify.

1 · Measure

We check how AI reaches and reads your site, how AI assistants describe your brand, and whether an agent can complete a buyer task.

2 · Diagnose

Every problem comes with its evidence: the page, the answer or the test step where it happened.

3 · Fix

Clear, ready-to-apply changes, reviewed by an engineer, delivered as code, tickets or a report your team can act on.

4 · Verify

We check again the same way. Technical fixes are confirmed directly; changes in AI answers are reported as trends.

Three questions, three kinds of evidence

What each part tells you.

What each part of the service covers
AspectFind + Understand
Readiness checks
Use
Agent journey test
AI answers
AI answer monitoring
The questionCan AI reach your pages and read the facts that matter?Can an AI agent complete a buyer task, like finding a price or reaching booking?How do AI assistants describe you, and who do they recommend instead?
What you seeEach issue with its page and a suggested fixA step-by-step replay showing where the agent got stuckTrends over time, with the answers as evidence
What it does not tell youWhat AI answers sayWhat AI answers sayExactly what each person sees: answers vary by person and time
Where to get itFree preview · full auditFull auditFull audit · Monitoring
The agent journey test

Watch an agent try to find the price and book.

A recorded test on our fictional demo practice, before and after the fixes. Same task, same site, two very different results.

Fictional demo treatments page before fixes: the price list is an image and the booking button is an unlabelled icon.
Before · stopped at step 3 The price list is a picture, so the agent finds no price.
Fictional demo treatments page after fixes: prices are in a table and a button reads Book an appointment.
After · task completed Prices are readable text and the booking link has a name.

Fictional business · recorded on our own demo site · Play the full replay

What the test shows you

  • Whether an agent can find your prices, reach booking or reach checkout.
  • The exact step where it got stuck, and why.
  • Results over repeated runs, so one lucky or unlucky run does not decide.
  • The same test again after the fixes, so you see the difference.

Safety commitments

  • Never submits a form, books, pays, logs in or enters personal data.
  • Runs on your site only after you confirm you own it.
  • Stops at logins, CAPTCHAs and bot challenges, and tells you.
  • Stays on your website: it cannot be sent to other sites.
Free preview vs full audit

What you get at each step.

Free preview compared with the full audit and monitoring
IncludedFree previewFull auditMonitoring
Readiness checks with a fix per issueKey public pagesYour key pages, in depthRe-checked monthly
Find / Understand / Use resultsYesYesYes, with trends
AI answer monitoringNot includedIncludedTracked monthly
Agent journey tests on your siteNot includedIncluded (Advanced)On request
Engineer-reviewed fix planSuggested fix per checkPrioritised planMonthly recommendations
ReportOn screen, share link, PDFFull PDF report + walkthroughMonthly report