The problem is not your reputation. It is your data.
ChatGPT, Perplexity, and Gemini are not search engines. They do not rank ten results and let the user choose. They pick two or three firms, recommend them by name, and ignore everyone else. The firms they pick are not necessarily the best — they are the best-structured. The ones that have told AI exactly who they are, what they do, and why they are credible.
TendorAI tracks AI recommendation data across 12,793 SRA, ICAEW, and FCA-registered firms in the UK. The data is unambiguous. Here is what it shows.
91% of UK Professional Services Firms Are Missing the Data AI Needs
AI cannot recommend what it cannot read — and right now, it cannot read most UK professional services websites.
TendorAI’s analysis of 8,625 SRA-registered solicitors found that 91% have no structured data markup on their website. No Schema.org. No machine-readable service information. No verified regulatory credentials in a format AI can parse. To ChatGPT, these firms are effectively anonymous.
The same pattern holds across verticals. Of 35,000+ ICAEW and ACCA-registered accountancy practices, fewer than 9% have structured data in place. Among FCA-registered mortgage advisers, the figure is lower still.
This is not a small gap. It is the difference between existing to AI and being invisible to it.
The 6 Data Gaps That Keep Firms Out of AI Recommendations
TendorAI’s weekly scans identify the same gaps repeatedly across every vertical. These are the six most common — ranked by frequency across all firms tracked.
No structured data markup on the website
Present in 91% of cases. Schema.org markup is the language AI speaks. Without it, your website is a document AI has to guess the meaning of rather than read directly. A firm with Schema markup installed is cited in AI responses 3.2 times more often than one without, according to TendorAI’s scan data.
Inconsistent NAP data
Name, address, and phone number appearing differently across your website, Google Business Profile, Companies House, and directory listings. AI cross-references these sources to verify an entity is real. Inconsistencies reduce confidence and reduce recommendation frequency. TendorAI’s data shows 40% of tracked firms have at least one NAP inconsistency across platforms.
No fee or pricing information
AI recommends specifics. When someone asks “how much does a conveyancing solicitor cost in Leeds,” AI looks for a source that answers that question directly. A firm with published fee ranges gets cited. A firm with “contact us for a quote” gets skipped. 78% of solicitor profiles in TendorAI’s dataset have no fee information whatsoever.
Regulatory credentials not machine-readable
Your SRA number, ICAEW registration, or FCA authorisation number may appear on your website — but if it is buried in a paragraph of text rather than marked up as a structured data field, AI cannot extract it. Without verified regulatory credentials, AI treats your firm as unconfirmed and reduces recommendation confidence accordingly.
No reviews or social proof in structured format
BrightLocal’s research shows 88% of consumers trust online reviews as much as personal recommendations. AI applies similar weighting. A firm with AggregateRating schema showing verified reviews scores significantly higher than one with no review data — even if reviews exist on Google but are not surfaced in structured format.
No fresh content answering the questions clients ask AI
AI recommendation is not a one-time fix. AI platforms crawl and update continuously. A firm that published a blog post in 2023 and nothing since is losing ground to firms publishing structured, FAQ-rich content monthly. The firms appearing consistently in AI responses are publishing content that directly answers the questions their clients ask — with specific UK pricing, regulatory context, and structured FAQ markup.
What a 28/100 Score Actually Means
The average UK professional services firm scores 28 out of 100 on TendorAI’s AI Visibility Score. At that level, AI knows your firm exists — but not enough to recommend it with confidence.
The score measures six components: schema installation, NAP consistency, fee transparency, regulatory credential verification, review data, and content freshness. A score of 28 typically means a firm has a Google Business Profile, appears in one or two directories, and has an SRA or ICAEW number somewhere on their website — but none of it is structured, verified, or connected in a way AI can use.
| AI Visibility Score | What It Means | Recommendation Likelihood |
|---|---|---|
| 0–30 | AI knows you exist, cannot verify you | Rarely recommended |
| 31–60 | Partial data — AI mentions but does not prioritise | Occasionally recommended |
| 61–80 | Strong structured data — AI recommends with context | Regularly recommended |
| 81–100 | Complete entity profile — AI recommends with confidence | Consistently recommended |
The firms scoring above 60 in TendorAI’s dataset have three things in common: schema markup installed on their own website, verified regulatory credentials in structured format, and at least one recent piece of content answering a specific client question.
Why Google Rankings Do Not Transfer to AI
A firm ranked on page one of Google for “accountant Manchester” can be completely absent from ChatGPT’s answer to the same question. This is not a bug — it is a fundamental difference in how the two systems work.
Google ranks pages. AI recommends entities. Google looks at backlinks, keyword relevance, and domain authority. AI looks at structured data, entity verification, and confidence signals. The two systems share some inputs — domain authority has some crossover — but they are not the same channel and optimising for one does not optimise for the other.
This is why firms with strong Google rankings are surprised to find themselves invisible in AI responses. And it is why firms with modest Google presence but complete structured data profiles are appearing in ChatGPT recommendations ahead of established competitors.
The Firms AI Does Recommend — What They Have in Common
TendorAI’s scan data across 10 major UK cities shows the firms appearing consistently in AI recommendations share a consistent profile:
- Schema.org markup installed and current
- ICAEW, SRA, or FCA number confirmed in structured format
- Fee ranges published for primary services
- Google Business Profile complete and consistent with website data
- At least 10 verified reviews with AggregateRating schema
- Content published within the last 90 days answering a specific client question
None of these firms are necessarily the largest or the best-known. Several are small practices with fewer than five staff. What they have is complete, structured, verified data — the exact signals AI needs to recommend with confidence.
Frequently Asked Questions
Why does AI recommend my competitor instead of me?
Your competitor almost certainly has structured data in place that you do not. Schema.org markup, verified regulatory credentials, published fee information, and consistent NAP data across platforms are the primary signals AI uses to decide who to recommend. TendorAI’s scan data shows firms with complete structured data profiles are cited 3.2 times more often than those without.
Does being on Google My Business help with AI recommendations?
Partially. A complete, consistent Google Business Profile is one signal AI uses — but it is not sufficient on its own. AI also needs structured data on your website, verified regulatory credentials, and fee information. A firm with a strong GBP but no schema markup on their website will still score in the 20–35 range on TendorAI’s AI Visibility Score.
How long does it take to start appearing in AI recommendations after fixing the gaps?
Structural fixes like schema installation typically take 4–8 weeks to be reflected in AI responses. Perplexity updates fastest — often within 2–3 weeks. ChatGPT takes longer. TendorAI’s tracking data shows firms that complete their structured data profile see AI mention frequency increase by an average of 340% within 90 days.
Is AI visibility only relevant for firms in large cities?
No — and smaller cities and towns often present a bigger opportunity. In major cities like London and Manchester, AI recommendation competition is higher. In smaller markets — Cardiff, Leicester, Cheltenham, Preston — fewer firms have structured data in place, which means a complete profile stands out more clearly.
What is the fastest way to improve my AI Visibility Score?
The single highest-impact action is installing Schema.org markup on your website with your regulatory credentials, services, and fee ranges structured correctly. This alone typically moves a score from the 20–30 range to 50–65. The free AEO report at tendorai.com/aeo-report shows your current score and the specific gaps holding it down — it takes 30 seconds and requires no login.
The data is not ambiguous. 91% of UK professional services firms are invisible to AI because they are missing structured data — not because they are unknown, unqualified, or lacking clients.
The firms AI recommends are not the biggest or the oldest. They are the best-structured. That is a gap that closes in weeks, not years.