How UK Law Firms Get Recommended by ChatGPT and AI Search

Someone with a redundancy dispute or a house purchase used to type "employment solicitor near me" and scan a list of links. Now they ask ChatGPT or Google's AI a full question and read a short written answer instead. That answer recommends two or three named firms, and everyone else has effectively vanished.

Almost everything written on this subject is aimed at the US market, and it skips the part that decides the outcome here: the way UK legal regulation and the trust signals around it feed into what a model is willing to say. So this guide sticks to what firms in England and Wales can actually check and fix, the order to do it in, and what to expect once you have. It is about marketing visibility, nothing more. It is not legal advice, and it is not SRA compliance advice.

How clients now ask AI about legal problems

Instead of scanning ten blue links, people now put whole questions to ChatGPT and Google's AI. Do they even need a solicitor for this, or can they handle it themselves? Who is the best firm for their kind of matter in their city? What is a fixed fee for conveyancing, and what will a divorce actually involve? The AI answers in full sentences, and those sentences name specific firms. If yours is not one of them, the client may never learn you exist.

These questions are conversational and oddly precise. "Best solicitor for conveyancing in Bristol." "Do I need a solicitor to contest a will?" "Employment solicitor in Leeds for an unfair dismissal, no win no fee." "How much does an uncontested divorce cost with a solicitor in Manchester?" Each one is a person part-way to instructing someone, using the AI to shortlist for them.

Two things set this apart from the old way of searching. The person often arrives having already worked out, in the same conversation, whether they have a case and roughly what it involves, so they are much closer to instructing a firm than a casual browser ever was. And the AI has done the shortlisting on their behalf, compressing a whole city's worth of firms down to a handful of names, each with a line of reasoning attached. That second point is the one that should keep you up at night.

Why AI is so selective about naming firms

AI assistants hand out very few local recommendations. SOCi's 2026 Local Visibility Index found that ChatGPT names only about 1.2% of the 350,000 business locations it analysed, against the 35.9% that turn up in Google's local three-pack (SOCi, 2026). An AI answer is a shortlist of two or three names, not a directory, and the overwhelming majority of firms never make it onto that list.

Meanwhile the audience for those shortlists is expanding fast. The same research found that 45% of consumers now use AI to find local businesses, up from just 6% in 2025 (SOCi, 2026). Ranking well on Google buys you no free pass here either. When Ahrefs ran 15,000 prompts through ChatGPT, Gemini and Copilot in 2025 and checked the citations against Google's rankings, only about 12% of the pages those assistants quoted also ranked in Google's top ten for the same question (Ahrefs, 2025). The traffic is not shrinking so much as moving: AI platforms sent 1.13 billion visits to the top 1,000 websites in June 2025, up 357% year on year (Similarweb via TechCrunch, 2025).

Why the ruthlessness? A results page can hedge its bets and list twenty options; a written answer cannot. The moment a model commits to naming a law firm, it puts its own credibility on the line, and legal services are exactly the kind of high-stakes, money-and-rights territory where a model is cautious. So it leans on evidence it can confirm from more than one source: regulation it can check, reviews it can read, and facts about you that agree across the web. If you want the mechanics in more depth, we go through them in our guide to how UK businesses get recommended by ChatGPT, and the sibling piece on how UK dental practices get recommended covers the same ground for a regulated healthcare market.

The UK trust signals that decide who gets named

When a model recommends a regulated professional, it wants evidence it can stand behind. For a law firm in England and Wales, that evidence has a very particular shape: regulatory registration and transparency, named solicitors you can look up, reviews on the platforms the models actually read, and content that proves you genuinely know your practice area. This is exactly the layer US advice leaves out, and it is where most British firms either win or disappear. Here are the signals that carry the most weight.

SRA regulation and transparency, made obvious

Solicitors in England and Wales are regulated by the Solicitors Regulation Authority, and anyone can verify a firm or an individual through the SRA's register or the Law Society's Find a Solicitor service. Make that verification effortless: display your SRA number and the SRA digital badge prominently, and make sure the details on your site match the register exactly. There is a happy overlap here with a rule you already have to follow. Under the SRA Transparency Rules, firms that publish the availability of certain services, conveyancing among them, must show price and service information and their complaints procedure, including how to complain to the Legal Ombudsman and the SRA (SRA, Transparency Rules). That is precisely the concrete, checkable content an AI reaches for, so compliance and visibility pull in the same direction. Firms in Scotland and Northern Ireland are regulated differently, by the Law Society of Scotland and the Law Society of Northern Ireland respectively, so use the equivalent signals there.

Named solicitors with real credentials, on your site

This is E-E-A-T in its most literal form: experience, expertise, authoritativeness and trust, expressed as real people a machine can verify. Give every fee earner a proper profile: full name, role, the year they qualified or their SRA status, the practice areas they actually handle, and notable experience. A page that shrugs and says "our expert team" hands a model nothing it can check. A named solicitor whose SRA record, website bio, LinkedIn and any directory listing all tell the same story is the opposite: a signal a machine can verify and lean on. Where a solicitor holds a recognised accreditation, such as membership of a Law Society accreditation scheme in their field, say so plainly and accurately.

Genuine reviews, on the platforms AI actually reads

Google reviews carry the most weight, simply because both Google's AI and ChatGPT draw on them. On top of that, the legal sector has its own platform: ReviewSolicitors, which the SRA lists among the digital comparison tools taking part in its voluntary code, alongside Trustpilot, The Law Superstore and reallymoving (SRA, customer reviews and comparison websites). A review that names the matter type, the solicitor and the town gives a model something concrete to work with, so volume helps but detail helps more. Reply to reviews like a professional, the critical ones included, and mind confidentiality when you do. And never buy or invent reviews: the platforms catch them, the SRA takes a dim view, and so do clients.

Practice-area pages with real depth

Generic firms lose here. A model asked for the best firm for a specific kind of matter is looking for evidence of genuine depth in that exact area, not a one-line services list. So give each significant practice area its own page, and make it substantial: what the process involves, how you charge, who leads the work, the sort of cases you handle. A firm with a thorough, plainly written employment page that sets out the fee options and names the solicitor who leads that team is exactly the concrete material a model reaches for when someone asks about unfair dismissal in your city. A rival with three sentences under an "Employment" heading is not.

Plain-English guides that answer real client questions

One of the most common things people ask AI is whether they need a solicitor at all, and for what. Firms that answer those questions honestly, in plain English, become the source the model quotes. Write guides around the real questions: do I need a solicitor for probate, what happens at each stage of a purchase, how is a financial settlement worked out on divorce. Put the question in the heading and answer it in the opening sentences. Keep it genuinely useful rather than a thin advert, and keep it accurate, because the SRA requires that anything you publish is not misleading. This is the single biggest opportunity most firms leave on the table.

Directory signals: Legal 500 and Chambers

Independent corroboration matters, and legal has strong, respected directories in the Legal 500 and Chambers and Partners. Both rank firms and individuals on independent research and client feedback, and a ranking is a third-party trust signal a model can weigh. You cannot buy a ranking, and you should never imply one you do not hold, but if you are ranked, make sure it is stated accurately on your site and profiles. Well-kept listings on the Law Society's Find a Solicitor and reputable legal directories add to the web of consistent, checkable facts about you.

Consistent name, address and phone everywhere

Your firm name, office address and phone number need to read identically across your website footer, your Google Business Profile, the SRA register, your directory listings and ReviewSolicitors. Models cross-reference these sources before they commit to a name. When your details contradict each other from one listing to the next, you become harder to verify, and the model quietly reaches for a competitor whose facts line up. Multi-office firms should be especially careful to keep each office distinct and consistent.

Schema markup, with an honest caveat

Structured data spells out your firm in a format machines read cleanly: LegalService or Attorney and LocalBusiness for the firm and its offices, and FAQ schema on your guides so a model can see the question-and-answer shape directly. It is worth having. Just be honest with yourself about what it buys you. Ahrefs tracked 1,885 pages that added schema and found no meaningful uplift in AI citations on any platform (Ahrefs, 2026). Schema helps machines understand you; it does not conjure citations out of thin air. Add it once the facts it describes are accurate, then move on and spend your energy on regulation, reviews and content.

A note on the SRA rules, kept straight

It is worth being clear, because the myth that solicitors "cannot really market" still floats around. The SRA does not ban marketing. For firms in England and Wales, the rules that matter most here are Principle 4, which requires you to act with honesty, and paragraph 8.9 of the Code of Conduct, which requires that publicity in relation to your practice is accurate and not misleading. There is a separate rule against making unsolicited approaches to members of the public. Everything in this article, accurate prices, honest solicitor profiles, genuine reviews, plain-English guides, sits comfortably inside those rules, and the Transparency Rules actively require some of it. The one hard line is honesty: do not overstate outcomes, do not claim rankings or accreditations you do not hold, and do not let an AI-friendly page drift into a misleading one. If you are unsure how a rule applies to your firm, check the SRA guidance or take compliance advice, because this is general information rather than a substitute for either.

A 10-minute self-test: what does AI say about you?

Open ChatGPT and Google and ask them the questions your own clients ask, using your practice areas and your city. Watch which firms get named, what the AI says about each one, and whether a single thing it says about you is wrong. Ten minutes is enough to see exactly where you stand.

1. In ChatGPT, ask: "best solicitor for [your main practice area] in [your city]", then "do I need a solicitor for [a common matter you handle]", then "how much does [that matter] cost with a solicitor in [your city]".
2. Repeat the same questions in Google and read the AI-generated answer at the top, not the links below it.
3. Add one question for each of your two highest-value practice areas.
4. Write down every firm named, and the reason the AI gives for naming it.
5. Check what, if anything, is said about your firm. Note any errors: wrong office address, a practice area you do not offer, a solicitor who has left still listed, an outdated fee.
6. Save the exact answers and date them, so you have a record to compare against in three months.

The answers shift between sessions and accounts, so treat any single run as a snapshot rather than a verdict. But if competitors keep getting named and you never do, the sections above will usually tell you where the gap is.

What to fix first, in order

Start with the signals AI can verify today: your Google Business Profile, contact details that agree with each other, your SRA number and transparency information, and named solicitors on your site. Reviews, practice-area depth and guides come next. Schema goes last. The logic is simple: put the trust signals that separate recommended firms from invisible ones at the front, and save the technical polish for once the foundations are actually in place.

1. Google Business Profile. Correct category, full list of practice areas, accurate hours, current photos, and a description that names your city and the work you do.
2. Contact detail consistency. Audit every listing that mentions your firm and make name, address and phone identical, per office.
3. SRA and transparency signals. Display your SRA number and digital badge, and make sure your price, service and complaints information meets the Transparency Rules where they apply to you.
4. Solicitor profiles. Named fee earners with roles, qualification details and the practice areas they actually handle.
5. Reviews. A steady, genuine flow of Google and ReviewSolicitors reviews, asked for consistently and responded to professionally.
6. Practice-area pages. Give each significant area real depth: process, fees, the people who lead it, the cases you take.
7. Plain-English guides. Answer the real questions clients ask, honestly and accurately, with the question in the heading.
8. Directories. Keep Find a Solicitor and any Legal 500 or Chambers listing accurate and up to date.
9. Schema. Add LegalService or Attorney, LocalBusiness and FAQ markup once everything it describes is accurate.

For the general method behind this sequence, see our UK guide to AI search optimisation.

How long it takes, honestly

Reckon on weeks before AI systems re-read your updated pages and profiles, and months before the answers clients actually see start to move. Nobody can guarantee you a ChatGPT recommendation. What sits firmly within your control is whether every checkable fact about your firm is accurate, consistent and easy for a machine to confirm.

AI answers are stitched together from a mix of live web results and older training data, which is why change arrives so unevenly. Fix your Google Business Profile and you can nudge answers within weeks. Build a genuine reputation through reviews and stronger practice-area content and it compounds quietly over months. So re-run the self-test every month with the same questions and hold on to your records; the trend tells you far more than any single answer ever will. And if a supplier promises you a named spot in ChatGPT answers by a fixed date, be sceptical, because no honest one can control that.

Questions people ask

These are the questions partners and practice managers put to us most often once the subject of AI visibility comes up. The answers are short and straight. And if your situation is more particular than any of them, the self-test above will almost always tell you more than general guidance can, because it shows you the exact answers your own clients are seeing.

Do people really use ChatGPT to find a solicitor?

Yes, and the shift has been fast. SOCi's 2026 research found 45% of consumers now use AI to find local businesses, up from 6% in 2025. Legal queries fit the pattern because people ask whether they even need a solicitor, what a matter will cost and who to trust before they make contact. The answers frequently name specific firms.

Do the SRA rules stop my firm from marketing to get named by AI?

No. The SRA does not ban marketing. For firms in England and Wales, Principle 4 requires honesty and paragraph 8.9 of the Code of Conduct requires that publicity is not misleading, and there is a separate rule against unsolicited approaches to the public. Publishing accurate prices, plain-English guides and honest solicitor profiles is squarely within the rules. In fact the Transparency Rules require firms offering certain services to publish price and complaints information anyway, which is the same content AI systems reward. This is general information, not compliance advice.

Will adding schema markup get my firm recommended?

Not by itself. Schema helps AI systems understand who you are, which practice areas you cover and where you are. But Ahrefs tracked 1,885 pages that added schema in 2026 and found no meaningful uplift in AI citations. Treat schema as useful plumbing, not a shortcut. Regulation, reviews and genuinely useful content do the heavy lifting.

How long before AI answers mention my firm?

Fixes to your website and profiles are typically re-read within weeks, but the recommendations clients see usually take months to shift, and some answers change slowly or not at all. No agency can promise a specific result. Be wary of anyone who guarantees a ChatGPT recommendation by a set date, because nobody controls what a model says.

Sources

• SOCi, Local Visibility Index (2026)
• Ahrefs, only 12% of AI-cited URLs rank in Google's top 10 for the original prompt (2025)
• Ahrefs, schema and AI citations study (2026)
• Similarweb via TechCrunch, AI referrals up 357% year over year, reaching 1.13B (2025)
• Solicitors Regulation Authority, Transparency Rules
• Solicitors Regulation Authority, Code of Conduct for Solicitors (Principle 4 and paragraph 8.9)
• Solicitors Regulation Authority, customer reviews and comparison websites
• The Law Society, Find a Solicitor
• ReviewSolicitors
• The Legal 500
• Chambers and Partners