More patients now ask an AI assistant before they search. "Where can I get ear wax removal near me?" "Which pharmacy in Hamilton does travel vaccines?" "Is there a private physio open on Saturdays in Paisley?" The answer comes back as a short list of names, and if a clinic isn't on it, the patient never sees its website.
I work on search for independent pharmacies, private clinics and dental practices in Scotland and the rest of the UK. This is the check I now run for every one of them, alongside the usual Google work. It takes about half an hour.
1. Ask the questions patients actually ask
I write down five or six questions in the patient's words, not the clinic's. A pharmacy might call it "microsuction ear care". The patient types "ear wax removal near me".
For each service I use three kinds of question:
- Service and place: "ear wax removal in Larkhall"
- Service with a condition: "ear wax removal open Saturday near Hamilton"
- Comparison: "best pharmacy for travel vaccines in Lanarkshire"
I run them in ChatGPT with search turned on, in Google (looking at the AI Overview when one appears), and on an iPhone with Siri. I note which businesses are named, in what order, and which sources each answer shows.
2. Look at what the answers cite
This is the useful part. The answers rarely invent anything. They pull from a small set of places, and once you see which ones, you know what to fix.
In my checks, the same sources come up again and again:
- The Google Business Profile: name, category, services, opening hours, reviews and the website link.
- The clinic's own service pages: especially pages that say what the service is, who it suits, the price and how to book.
- Reviews: on Google, and sometimes on sites like Trustpilot or Doctify.
- Directories and NHS listings: used to confirm the name, address and phone number.
- Bing: ChatGPT's search leans on Bing results, so a clinic missing from Bing Places is often missing from ChatGPT too.
3. Fix the five things that change the answer
When a clinic is missing, it is almost always one of these.
The Google Business Profile is half filled. No services listed, the wrong main category, old opening hours. Fill in every field. Add each service with a one-line description. Check the hours match the website.
There is no page for the service. A single "Services" page with a list of names gives an AI nothing to quote. Each service needs its own page that answers the patient's question in the first lines: what it is, who it is for, what it costs and how to book.
The details don't match. If the website says Unit 3, Google says Unit 3A and a directory has the old phone number, the answers hedge or leave the clinic out. Make the name, address and phone number identical everywhere.
Reviews are thin or unanswered. Recent reviews that mention the service by name help. Replying to them, within professional guidance, helps too.
There is no clear booking route. Answers like to point patients somewhere they can act. A booking link on the Google profile and on each service page makes that easy.
4. Track it
Most AI answers don't show up as referrals in analytics, so I keep it simple:
- Re-run the same questions every month and note whether the clinic appears.
- Add "How did you find us?" to the booking form, with "ChatGPT or another AI assistant" as an option.
- Watch Google Search Console for the service pages getting more impressions, since the same fixes help normal search too.
The short version
AI answers aren't a separate channel with their own tricks. They reward the same basics as local search: a complete Google profile, one clear page per service, consistent details and real reviews. The difference is that the gap between a clinic that has done the basics and one that hasn't is now much starker, because the patient only sees three names.
I've written up how this fits with the rest of our work on AI search visibility for clinics and pharmacies.

Top comments (0)