Fictional probate practice · scripted browser checks · illustrative summary layout
Two small changes.
Inspected separately.
A visitor needs a working route and clear information. These examples show exactly what changed on one fictional website, one change at a time.
Repair the consultation link
Before: the main consultation link led to an unavailable page. After: that one link points to the existing consultation page. The summary and all other content stay the same.


Before: 6 / 6 checks completed; 0 excluded. After: 6 / 6 checks completed; 0 excluded. Three repetitions at each of two screen sizes per condition.
Fictional example. Same configuration within this comparison; only the stated change. Scripted browser checks, no AI model.
Publish the missing details
Before: the summary did not state consultation length, the consultation fee or how later work is priced. After simulated approval, those three details are added. The contact route and all other pages stay the same.


Before: 6 / 6 checks completed; 0 excluded. After: 6 / 6 checks completed; 0 excluded. Three repetitions at each of two screen sizes per condition.
Fictional example. Same configuration within this comparison; only the stated change. Scripted browser checks, no AI model.
What this demonstrates
The repaired link reached an existing consultation page. Publishing the approved example details made them explicitly available on the summary page. Those are the observed changes.
No AI model or consumer assistant ran. The checks did not test organic discovery, extraction by an AI product, recommendations or use of llms.txt. Repeated checks at two screen sizes are smoke tests, not a statistical sample.
Explore the complete fictional website →Method & evidence
Recorded total: 24 / 24 completed browser observations, 0 excluded. Execution completion is separate from the observed route or information result. The browser ran in a separate process with page JavaScript disabled and an explicit local-only website.
Both conditions use the same frozen configuration within each comparison. Structured data is unchanged: 0 JSON-LD blocks in either condition. This does not validate structured data. Exact source changes, configuration, events and captures are retained in the downloads.
| Fictional fact | Before | After |
|---|---|---|
| Initial consultation length | 0 / 6 | 6 / 6 |
| Fixed consultation fee | 0 / 6 | 6 / 6 |
| Later work quoted separately | 0 / 6 | 6 / 6 |
The fictional facts were approved in a simulation: a 30-minute initial consultation, a fixed consultation fee of £120 and a separate quote for later work. These are not BrowseProof prices or real legal-service terms.
Only the footer link uses exact supported plugin markup. The summary layout is illustrative, rather than the standard generated-page layout. This is not a fresh WordPress installation. Supporting text files are unchanged in each comparison and are not read by these browser journeys.
- Recorded comparison summary and configuration hashes
- All observations (CSV)
- Exactly what changed in each comparison
- Full fictional evidence: sources, events and original captures (ZIP)
- Evidence file manifest
Historical technical downloads
Earlier, separate rehearsals are preserved unchanged. They do not establish results for this example.
- 7 October: fictional WordPress rehearsal — simulated approvals, 10/10 required files matched. Provenance.
- 11 October: earlier scripted audit report — contact route and structured data changed together, so their effects cannot be separated. Method record.
Investigate a task on your website
Agree the question and method before running a check.
Discuss a website audit →