Key Takeaways
- A 173-site audit conducted by Qbiz Leads AI, of UK local service websites found that 94%+ pass fundamental technical checks like crawlability and mobile rendering - but only around 31% carry the answer-shaped content AI systems actually read.
- The three most-failed signals are all about explanation: FAQ content (absent on 68% of sites), process descriptions (missing on 67%), and llms.txt files (missing on 69%).
- Ranking on Google and being readable by AI are two separate problems - a site can sit on page one and still be invisible to AI answer engines.
- The fix does not require a website rebuild - it is a content and structure problem, and most sites are already one or two additions away from closing the gap.
Most UK local service businesses are not invisible online. They have websites that load, rank, and look the part. But a fresh look at 173 of those sites reveals something worth paying attention to: the technical side of being online and the content side of being understood by AI are two very different things - and right now, most sites are winning the first battle while quietly losing the second.
Present Online, but Unreadable by AI
There is a version of online visibility that most local businesses have already achieved. The website loads. Google can crawl it. It works on a phone. The business name, address, and phone number are on the page. By every traditional measure, the site is doing its job.
But search has changed. When someone asks ChatGPT to recommend a solicitor in Manchester, or uses Google AI Overviews to find a local dentist before clicking anything, those platforms are not browsing a list of ranked pages - they are reading sources and pulling direct answers. If a website does not carry the kind of structured, answer-ready content those systems are looking for, it simply does not make the cut, regardless of how well it ranks.
Q Biz Leads AI audited 173 UK local service websites across four sectors and five cities, scoring each against ten binary readiness signals. Their AI optimisation services are built around exactly this kind of structural gap - but the audit findings stand on their own: the visibility problem most local businesses face is not about being absent from the web. It is about being under-explained on it.
What the 173-Site Audit By QBiz leads AI Found
The audit covered plumbers, solicitors, dental practices, and accountants across London, Manchester, Birmingham, Edinburgh, and Bristol. Each site was scored on ten signals and placed into one of three bands: Not Ready (0-3), Partially Ready (4-6), or Ready (7-10).
The headline result: 120 of 173 sites (69.4%) scored as Ready. The mean score across all sites was 6.92 - just below the Ready threshold of 7. On the surface, that sounds encouraging. But the signal-by-signal breakdown tells a more pointed story.
Technical Basics: 94%+ Pass Crawlability, Mobile and Entity Signals
The technical foundations are largely solid. Crawlability passed on 94.2% of sites, mobile rendering on 93.6%, and entity consistency - meaning both a phone number and an address were present on the page - on 92.5%. Internal linking, service definitions, and Schema markup all passed on more than 79% of sites. These are the signals that have mattered for traditional SEO, and most sites have them covered.
The Answer Layer: Roughly Two-Thirds Fail the Signals AI Reads Most
Then there is the other end of the table. The three signals with the lowest pass rates are not obscure technical requirements - they are about whether the site actually explains what the business does:
- llms.txt present: passed on just 30.6% of sites (69.4% fail)
- FAQ content: passed on 31.8% of sites (68.2% fail)
- Process explanation: passed on 32.9% of sites (67.1% fail)
That is a clean split. The signals AI systems rely on most heavily for generating direct answers are the ones most sites have simply never added.
Ranking and AI Readability Are Two Different Problems
This is the part that catches many business owners off guard. A site can appear on page one of Google and still be effectively invisible to the AI layer sitting above the results. Traditional SEO - earning backlinks, matching keywords, loading fast - was built for a ranked-list model of search. AI answer engines work differently. They read available sources, extract structured information, and construct a direct response.
Research into zero-click search behaviour shows that a growing share of queries are now resolved directly within AI-powered summaries, before a user clicks any result. Separately, data from BrightEdge indicates that 62% of brands lose visibility in AI environments compared to traditional search. If a website does not support that kind of extraction - no FAQ blocks, no process descriptions, no machine-readable guide - the answer engine fills the gap using whoever's content it can read. Often, that means a competitor.
The concept gaining traction in response to this shift is Answer Engine Optimisation (AEO): structuring content so AI platforms can find, read, and cite it. It is a separate layer of work from SEO, focused on explanation rather than ranking.
The Three Answer Signals Most Sites Skip
FAQ Content: Absent on 68% of Sites
FAQ sections are among the most extractable content formats on any website. A well-written question-and-answer pair is a self-contained unit - exactly the kind of thing an AI system can lift and use directly in a generated answer. Sites with structured FAQ content are more likely to appear in People Also Ask boxes and AI summaries, because these formats directly address user intent.
Yet 68% of audited sites had none. Not thin FAQs - none at all. The questions a customer asks before booking a plumber or a solicitor exist whether the site addresses them or not. When the site does not, the AI looks elsewhere.
Process Explanation: Missing on 67% of Sites
Two in three sites gave no description of what actually happens when a customer makes contact. No how-it-works section. No what-to-expect walkthrough. No step-by-step description of the journey from enquiry to completed job.
This matters both for human readers deciding whether to get in touch, and for AI systems trying to distinguish one local service business from another. A process description is answer-ready content by nature - structured, sequential, and easy to extract.
llms.txt: The Single Most-Failed Signal at 69%
An llms.txt file sits at the root of a website and acts as a plain-text guide for AI crawlers - similar in concept to how robots.txt directs traditional search bots, but specifically aimed at large language models. It tells AI systems what the site contains and how to navigate it.
It was present on just 30.6% of audited sites, making it the single most-failed signal in the study. The audit treats it as an emerging signal rather than a proven requirement - but with fewer than one in three sites carrying one, adding it today is a genuine differentiator rather than a catch-up move.
How Sectors and Cities Compare
Solicitors Lead on Mean Score; Dentists Lead on Ready Share
Sector differences exist but are modest. Solicitors posted the highest mean readiness score (7.13) and had no sites in the Not Ready band. Dental practices achieved the highest Ready share at 81.6% - more than four in five dental sites cleared the threshold. Accountants had the lowest Ready share at 58.7%, held back by a large group of 17 sites that passed the technical basics but fell short on content and structure signals. Plumbers had the lowest mean score (6.67), partly due to fetch errors and a sizeable Partially Ready group. Every sector shows the same underlying shape: strong basics, thin answer content.
Manchester Strongest, Edinburgh Weakest Across Five Cities
Manchester led on the all-rows mean score (7.50) with an 87.5% Ready share and zero fetch errors. Edinburgh sat at the other end with the lowest all-rows mean (6.26) and the lowest successful-fetch mean (6.80) of any city - meaning its lower scores are not just a data artefact. London illustrates why the audit reports two calculation bases: five fetch errors dragged its all-rows mean to 6.67, but on the successful-only basis it jumped to 7.63, the highest of any city. The signal pattern - solid basics, weak answers - holds across all five cities regardless.
Ready vs. Partially Ready: What Each Looks Like
The tier labels can feel abstract, so it helps to see the actual shape of a site in each band. A representative Ready plumbing site in London scored 8 out of 10: it passed schema, headings, service definitions, entity consistency, crawlability, internal linking, mobile render, and llms.txt - falling short only on FAQ content and process explanation. One answer-layer edit away from a perfect score.
A Partially Ready plumbing site from the same city scored 6: it had the technical basics in place but missed schema, FAQ content, process explanation, and llms.txt entirely. Foundations solid, explanation absent. Not Ready sites - just 4 of the 173 audited - are a genuinely different problem. One accountancy site passed only FAQ content, crawlability, and mobile render, missing everything else. That is a thin web presence overall, not just an explanation gap, and the two situations call for different fixes.
The Fix Is Content, Not a Rebuild
The encouraging finding buried inside these numbers is that most local service sites are already closer than they think. Two in three are technically Ready. The vast majority that fall short are not missing the foundations - they are missing the explanation layer on top of them. That means the path forward does not involve redesigning a website or migrating to a new platform.
A quick self-audit takes minutes: open the homepage and a main service page, then check whether a first-time visitor could find a direct answer to a common question, a described process, a service list in plain words, and contact details on the page. Then check whether the site returns anything at yourdomain.co.uk/llms.txt. If several of those are missing, the site sits in the same position as most of the 173 audited - present, competent, and under-explained. The fixes to prioritise:
- Add a genuine FAQ section - written around the questions customers actually ask before booking, answered in plain, direct language
- Write a process or how-it-works description - what happens from first contact to finished job, in clear steps
- Consider an llms.txt file - a simple root-level guide for AI crawlers; uncommon enough today to be a real differentiator
- Tighten the structure - clean heading hierarchy, Schema markup, and consistent service-definition language throughout
The AI visibility gap is a content gap. And content gaps, unlike technical ones, do not require a developer to close.
Q Biz Leads AI helps UK local service businesses close exactly this gap - visit qbizleadsai.com to learn how they approach the space between ranking and being genuinely readable by AI systems.