A Sixty-Minute Relevance MOT for UK SMEs Losing Clicks to AI Answers

AI Overviews, Copilot answers and chat-style search are no longer a side experiment. For many UK small businesses they are becoming the first place a prospective customer encounters a brand — or, more often, encounters a summary that names competitors instead. Rankings can remain stable while clicks fall. That is why a relevance review, rather than a rankings report alone, is now a sensible monthly discipline.

Why sixty minutes is enough to find the leak

A full SEO audit can take weeks. A relevance review is narrower. It asks whether the business is easy for both people and AI-mediated search systems to understand, trust and choose. The aim is not to produce a perfect score but to identify where relevance is leaking: an unclear service page, an inconsistent address, a blog post that still refers to a former location, or analytics that cannot see enquiries that begin in an AI answer.

Set aside one hour. Use a spreadsheet with five columns: area, what you checked, evidence, score from 0 to 2, and the next action. A total of 10 is possible. Anything below 7 deserves attention within the month; anything below 4 should be treated as a commercial risk, not a marketing nicety.

Step one: test query intent against real customer language

Start with ten queries that matter commercially. Include a broad one such as 'accountant for contractors', a local one such as 'accountant Bristol city centre', and a problem-led one such as 'how to switch accountants mid-year'. For each, note whether the page that ranks or is cited actually answers the question in the first two paragraphs. AI summaries reward clear, extractable answers. A page that begins with a company history before reaching the answer is less likely to be quoted.

UK example: a Manchester heating engineer may rank for 'boiler repair Manchester', but the page may not explain emergency call-out cut-off times, landlord certificates or typical price ranges. Those omissions create a relevance gap that a competitor with a plainer page can fill.

Step two: make the entity unambiguous

Search systems need to know who the business is, what it does, where it operates and how it differs. Check the homepage, about page and primary service pages for consistent trading name, legal entity name, address, phone number and service area. If the business trades as a different name from its Companies House registration, say so clearly. Add a short, factual description that a machine can parse and a customer can understand.

Score 2 if the name, category, location and core service are stated plainly on the homepage and echoed on key service pages. Score 1 if some details are present but inconsistent. Score 0 if the site relies on logos, slogans or vague phrases such as 'solutions for every need'.

Step three: audit local citations and profile data

For local businesses, relevance is often won or lost in the details. Check Google Business Profile, Bing Places, Apple Business Connect and the main UK directories used in the sector. Look for old addresses, former phone numbers, incorrect opening hours and category labels that no longer fit. A Cumbrian guest house that still lists a pre-pandemic breakfast time is not merely untidy; it is feeding conflicting facts into systems that synthesise answers.

Also check the website's contact page, footer and schema markup. The same address should appear in the same format. If the business has moved, add a visible note about the former location only where it helps customers, not as a permanent archive of confusion.

Step four: test freshness against 2026 expectations

Freshness is not about changing dates for their own sake. It is about whether the page reflects current prices, regulations, opening hours, service areas and terminology. A Bristol accountancy page that still refers to Making Tax Digital as a future change, or a Yorkshire bakery that lists a discontinued product as seasonal, will feel less reliable than a page updated this year.

Score 2 if the top five commercial pages show a recent, substantive update. Score 1 if dates are current but the content is unchanged. Score 0 if the page contains references that are clearly out of date. Use a visible 'last reviewed' date only when it is true.

Step five: confirm measurement catches AI-assisted journeys

Traditional analytics often undercount journeys that begin with an AI summary and end with a branded search or a direct visit. Check whether enquiry forms, calls, WhatsApp links and booking tools have conversion tracking. Ask new customers how they found the business, and include a simple source question in the form. Review branded search volume alongside non-branded rankings. If clicks are falling but enquiries are steady, the issue may be measurement; if both are falling, relevance is the more likely suspect.

Score and prioritise

Use the five areas above, each scored 0-2. Add any evidence notes. Then rank actions by commercial impact and effort. A weak entity description on the homepage may take an hour to fix; a full content refresh may take a month. In the first seven days, correct factual inconsistencies, add plain answers to the top three commercial queries, update the most visible service page, and fix conversion tracking on the primary enquiry route.

The next step after the hour

Repeat the review quarterly. Keep the score sheet, because progress is easier to see when the criteria stay the same. The goal is not to chase every AI feature or rewrite every page. It is to remove the small points of friction that make a competent UK business look less relevant than it is.