Example report. Real test, business anonymised.
Clinic A: a private physiotherapy clinic in a UK city
Tested with Claude using live web search, location set to the city the clinic is in, 10 patient-style questions, single run. A business only counts if it was named in the answer text, not just found in the pages the assistant read.
- Named in 1 of 10 answers, even though it has a strong Google rating.
- Most-named competitors: two clinics were each named in 6 of 10, and two more in 5 of 10.
- Found but not named: for one more question the assistant read the clinic website, then recommended other clinics.
| Question asked | Clinic A named? |
|---|---|
| best physio in [city] for a sports injury | Yes |
| physiotherapist in [area] for a running injury | Nosite read, not named |
| private physiotherapist near [area] | No |
| who should I see for back pain in [city] | No |
| physio in [area] for a knee injury | No |
| best rated private physio for shoulder pain | No |
| sports massage and physio in the city centre | No |
| NHS or private self-referral | No |
| post-surgery rehab in [county] | No |
| neck and posture physio | No |
Likely reasons I found
- The assistant only saw Clinic A as a sports injury clinic, so it skipped it for back, neck and post-surgery questions.
- Competitors had separate pages for each neighbourhood and each condition. The assistant cited those pages directly.
- Clinic A was missing from several of the directories the assistant read, such as healthcare listing and booking sites.
What I would fix first
Condition and area pages on the site, a fuller Google Business Profile with every service listed, and profiles on the two or three directories the assistant kept citing.
One run, one assistant, one day. AI answers change between runs and over time, so a snapshot is a picture, not a verdict.