UK Law Firms in AI Search: Not One of 16 Firms Reached Moderate Presence
Joshua Etheridge
Not one of 16 audited UK solicitor firms reached even a Moderate AIO presence band when tested across approximately 50 real consumer AI prompts.
Every firm was named at least once, so the story is not that law firms are invisible. The story is that appearing occasionally is not the same as being recommended consistently, and on that measure, the entire sample fell short.
What We Measured and How
On 27 June 2026, AIO ran its live audit pipeline against 16 real UK solicitor firms drawn from four cities: Manchester, Leeds, Bristol, and Nottingham. Each firm faced the same battery of approximately 50 real prompts put to consumer ChatGPT and consumer Google Gemini on that date. We recorded, for each firm, in how many of those prompts the firm was actually named by the AI.
The headline metric is a raw count: named in N of 50 prompts. We do not publish weighted scores or proprietary indices here. The AIO presence band, our audit's own label, is assigned based on that count and reflects how consistently a firm surfaces across the prompt set.
One important limitation: these 16 firms were drawn from public legal directories and search results. That means the sample skews toward firms that already have some web presence. The true population of UK solicitor firms is unlikely to perform stronger than what we measured here.
The Core Finding
The average firm in this audit was named in about 15 of 50 prompts. The most-named firm, Slater Heelis in Manchester, was named in 25 of 50 prompts, and even that result placed it only in the Weak band. The least-named firms, MSD Solicitors and JWP Solicitors, were each named in 6 of 50 prompts. (One firm's battery, MSD Solicitors, generated 51 prompts rather than 50; for consistency we report all counts out of 50.)
No firm in the sample reached Moderate. No firm reached Strong. The band distribution tells the full picture.
AIO Band Distribution: All 16 Firms
| AIO Presence Band | Number of Firms | What It Means |
|---|---|---|
| Strong | 0 of 16 | Named consistently across the prompt set; the AI treats the firm as a go-to recommendation |
| Moderate | 0 of 16 | Named reliably across several prompt types; some topical authority established |
| Weak | 7 of 16 | Named in some prompts but not enough to be considered a consistent AI recommendation |
| No presence | 9 of 16 | Named rarely; the AI does not associate the firm with the practice areas or locations tested |
One note on how the bands are assigned. A firm's band comes from AIO's weighted presence score, which accounts for which prompts a firm was named in, not only how many. That is why a firm named in slightly more prompts can still sit in a lower band than one named in slightly fewer.
Full Per-Firm Results
| Firm | City | Named in (of 50 prompts) | AIO Band |
|---|---|---|---|
| Slater Heelis | Manchester | 25 of 50 | Weak |
| Smalleys Solicitors | Nottingham | 23 of 50 | Weak |
| Rothera Bray Solicitors | Nottingham | 22 of 50 | Weak |
| Kuits Solicitors | Manchester | 21 of 50 | Weak |
| Winston Solicitors | Leeds | 20 of 50 | Weak |
| Sills & Betteridge Solicitors | Nottingham | 20 of 50 | Weak |
| Stowe Family Law | Bristol | 17 of 50 | Weak |
| Buckles Solicitors | Bristol | 16 of 50 | No presence |
| Morrish Solicitors | Leeds | 15 of 50 | No presence |
| Fieldings Porter Solicitors | Manchester | 12 of 50 | No presence |
| Lyons Bowe Solicitors | Bristol | 11 of 50 | No presence |
| Brabners | Leeds | 10 of 50 | No presence |
| Michelmores | Bristol | 10 of 50 | No presence |
| Banner Jones Solicitors | Nottingham | 7 of 50 | No presence |
| MSD Solicitors | Manchester | 6 of 50 | No presence |
| JWP Solicitors | Leeds | 6 of 50 | No presence |
Why Appearing Once Is Not Enough
The tempting interpretation of this data is reassuring: all 16 firms appeared at least once, so AI tools do know they exist. That framing misses what actually matters to a prospective client using an AI assistant to choose a solicitor.
When a person asks ChatGPT or Google Gemini a question like "which solicitor in Leeds handles employment disputes for individuals", the AI produces a short list. Firms that are named in only 6 or 10 of 50 varied prompts are not on that short list reliably. They surface when the question is phrased in a narrow way that happens to match what the AI has absorbed about them. Change the phrasing slightly, and they disappear.
Slater Heelis, the top performer at 25 of 50 prompts, still missed roughly half the prompt variations we tested. Even the best result here is fragile coverage, not authority.
"Appearing in AI search is binary for most firms right now: they exist in the training data, but they have not given AI tools enough structured, consistent, authoritative signals to be recommended across a range of real questions."
What Drives the Gap Between Weak and No Presence
This audit measured correlations in output, not model internals. We cannot state with certainty why one firm surfaces more than another. What we can observe is that the firms landing in the Weak band tend to have longer public digital footprints, more structured content on practice-area pages, and more third-party mentions in legal directories and editorial coverage. That is a measured correlation from the audit output, not a claim about how ChatGPT or Gemini work internally.
The firms in the No presence band are not obscure businesses. Several are well-established regional practices with strong traditional reputations. The gap between traditional reputation and AI citation authority is one of the clearest findings from this audit.
The Business Case: An AIO Estimate
We want to be transparent about what is measured and what is modelled. The prompt counts above are measured. Any revenue figure attached to AI visibility is an estimate, not a measured fact.
AIO estimate, with the load-bearing assumption surfaced: If a mid-size regional solicitor firm converts one additional instructed matter per month that originated from an AI-assisted search, and if the average matter value for that firm is £3,000 (a figure that will vary widely by firm and practice area), the annual uplift is £36,000 in gross fees. We have no data on how many clients currently choose a solicitor based on an AI recommendation. That number is likely low today and growing. This estimate is included only to illustrate the order of magnitude; it is not a projected return on any specific investment.
What This Audit Does Not Tell Us
This is a focused audit of 16 firms across 4 cities on a single date. It is not a national census, and it makes no claim to represent all UK law firms. The prompt battery covers a defined set of practice areas and geographic queries; different prompts would produce different counts. The results reflect ChatGPT and Google Gemini as consumer products tested on 27 June 2026, not any specific underlying model version. Results from these tools change over time as their training and retrieval systems are updated.
Frequently Asked Questions
What does the AIO presence band mean for a law firm?
The AIO presence band is the label our audit assigns based on how many of approximately 50 real consumer prompts caused ChatGPT or Google Gemini to name a given firm. Weak means the firm was named in some prompts but not consistently enough to be considered a reliable AI recommendation. No presence means the firm was named rarely across the prompt set. Moderate and Strong bands indicate progressively more consistent citation, and no firm in this 16-firm audit reached either of those bands.
Were any firms completely invisible to AI tools?
No. Every one of the 16 firms audited was named in at least one prompt. The finding is not invisibility. It is shallow presence: being named occasionally is very different from being recommended consistently when a prospective client asks a relevant question.
Why was the sample limited to 16 firms across 4 cities?
This audit was a focused measurement exercise, not a national survey. Sixteen firms allowed us to run the full prompt battery rigorously and publish verifiable per-firm results. The firms were drawn from public legal directories and search results, which means the sample already skews toward firms with some existing web presence. The true population of UK solicitor firms is unlikely to perform stronger than what we measured here.
Does a higher named-in count prove the AI prefers that firm?
No. The counts reflect measured correlations in AI output across a defined prompt set on a specific date. We are not making any claim about model internals, training data weightings, or deliberate AI preferences. A higher count means the firm was named more often across the prompts we tested, which is useful information about current AI output but is not a guarantee of future performance.
How quickly can AI presence change?
AI tools like ChatGPT and Google Gemini update their underlying systems periodically. A firm's named-in count from this audit reflects a snapshot taken on 27 June 2026. Building more consistent AI citation authority typically requires sustained content and authority work over months rather than days, but AI tool updates can shift results in either direction at any time.
How can my firm find out where it stands?
AIO runs the same live audit pipeline used in this study against individual firms. You can request a free audit for your firm at aio.io. The audit report shows your named-in count across the prompt battery, your AIO presence band, and which practice area and location queries are driving or missing citations.