How clients find a local firm now: city searches and AI answers

New clients find a local firm in one of two ways now: they search the service and the place, a dentist in Leeds, a solicitor in Manchester, or they ask an AI assistant to recommend one. Both moments are decided before anyone rings your office, by whether a page exists for that exact search and whether it says something an answer engine can lift and repeat.

For clinics and professional firms, the second route is no longer marginal. The habit that decides where a new patient or client goes has moved twice in two years: first from the list of blue links to the answer at the top of the page, and then from the search engine to the assistant. Local visibility now means being present, and quotable, in both places.

Less and less. In 2024, 58.5% of Google searches in the United States ended without a click (SparkToro and Datos, 2024). The results page is increasingly the destination: the map pack, the ratings, the extracted answer. Ranking still matters, but the searcher often gets what they need without visiting anyone, which means the text the engine extracts is doing the selling.

The zero-click study is public (SparkToro and Datos, 2024), and the move to assistants is faster still. In BrightLocal's Local Consumer Review Survey, 45% of US consumers said they had used AI tools for local business recommendations in the past twelve months, up from 6% the year before (BrightLocal, 2026). The survey is American, but the direction of travel is the point. When someone asks an assistant for a good private dentist in their city, it names two or three practices and explains why. Either your firm is in that answer or it does not exist for that searcher.

None of this retires the basics. A firm that serves a county still wins or loses one search at a time, and each of those searches is a service plus a place. The dedicated page for that combination remains the thing that gets found, read, and quoted. We covered how to build these pages at scale in our guide to programmatic local SEO; the short version is that each one must carry genuine local information, not a place name swapped into a template.

What has changed is the readership. A city page is now read by people, by crawlers, and by language models assembling an answer. An assistant recommending solicitors in Bristol draws on the same local pages, directories, and reviews a careful person would check, then compresses them into three names and a sentence each. The page that is easiest to quote accurately is the page most likely to be named. That is the joint work of SEO and GEO: findable to the engine, citable in the answer.

This is now measurable rather than folklore. The first controlled study of generative engine optimisation tested content changes across thousands of queries and found that the right ones can raise a site's visibility in generative answers by up to 40%, with clear statistics, quotations, and cited sources among the strongest levers (Aggarwal et al., 2024). Fluent, specific, well-sourced writing is not a courtesy; it is how a model decides whom to quote.

The study is public (Aggarwal et al., 2024). For a clinic or professional firm, a city page earns its place in an answer the way a good reference earns a citation. Define your terms once and plainly, the way our glossary does for ours. Use the questions clients actually ask as headings, and answer each fully in the paragraph beneath, so the passage stands alone when lifted out. In practice, each city page should carry:

• One self-contained paragraph at the top that answers the head question outright: what you do, where you do it, who it is for, and how to start.
• Genuine local detail: the areas you cover, where patients or clients travel from, and anything specific about practising in that city.
• Question-shaped headings with complete answers beneath them, written to survive being quoted without the rest of the page.
• Figures with named sources and years rather than unattributed claims; answer engines prefer numbers they can trace.
• Identical firm details everywhere they appear: name, address, opening hours, and services, consistent across the page, your profiles, and the directories.

There is a quiet failure mode in local search, and it sits after the ranking. The page works, the enquiry arrives, and nobody answers it. In one audit of 433 B2B companies, 55% took more than five working days to reply to an enquiry, or never replied at all (Drift lead response survey, 2017). And the local searches that matter most often happen when the office is dark: in health and beauty, almost half of salon and spa bookings are made outside opening hours (Phorest).

A dental practice that wins the evening search for its city and then lets the enquiry sit until morning has paid for visibility and donated the patient to whoever answered first. This is why we treat the page and the response as one system: the page captures the enquiry, and the AI employees behind it answer, qualify, and book it into the diary at the hour it arrives.

Start narrow. Pick the cities where one new client covers a month of effort, and build or deepen the page for each with real local substance. Make the firm's details identical everywhere they appear. Rewrite the top of each page so the first paragraph answers the searcher outright, then add the questions clients ask and complete answers beneath them. Finally, wire the response, so the enquiries the pages produce are answered right away rather than days later.

None of this requires guesswork on budget: the pricing is public, and a thirty-minute call will map which cities are worth the work for your firm, and what being named in the answer should return.