ChatGPT has become a meaningful entry point in UK legal buyer journeys. Sceptical buyers ask it before they search Google. Younger buyers default to it. Partners may not see this in their analytics because the queries never reach the firm is website unless ChatGPT cites the firm.
For the same query, ChatGPT recommends specific firms over others, consistently. This is not random. The citation patterns are learnable, and the firms being recommended are doing identifiable things the others are not.
How ChatGPT actually answers UK legal queries
ChatGPT does not browse the web in real time on every query (it sometimes does, depending on the tier and the query, but not by default). For most queries about specific firms or recommendations, it answers from training data supplemented by browse-the-web calls when the model decides the question requires fresh information.
The training data is dominated by sources ChatGPT was trained on: the SRA register, ReviewSolicitors, Chambers and Legal 500, Wikipedia entries for major firms, and legal news sites. For firms that appear in those sources with substantive, structured information, ChatGPT has a model representation that includes them. Firms that exist mostly as a website with no third-party presence are largely invisible.
The four signals ChatGPT weights most
First, SRA register presence. ChatGPT consistently cross-references the SRA register to verify a firm exists and is regulated. Firms that are mentioned in pages that link to or reference their SRA registration number do better than firms that do not.
Second, ReviewSolicitors. The platform is structured for AI ingestion (named solicitors, verified reviews, practice area data) and ChatGPT treats it as a primary signal for consumer-facing legal queries. A firm with 40+ recent ReviewSolicitors reviews is going to outperform a firm with a similar website but no ReviewSolicitors presence.
Third, named, credentialled authors. ChatGPT models legal expertise as a property of people, not firms. A firm with practice area pages by-lined by named solicitors with clear credentials gets recommended more than a firm with anonymous “the team” attribution.
Fourth, content depth on procedural questions. The queries that drive most ChatGPT legal recommendations are not “best [firm]” — they are “how do I [legal procedure]”. Firms whose content actually answers those questions, with named solicitors as authors, get cited as sources when ChatGPT generates its answer.
Why most UK law firms are invisible to ChatGPT
Three structural reasons. Most firm websites lack named authors on content pages. Most firms have no ReviewSolicitors presence or have not invested in active review acquisition there. Most firms have no Person schema, so AI assistants have no machine-readable way to link content to a named, credentialled author.
The result is that ChatGPT, when asked “best probate solicitor in [city]”, defaults to the firms that have credentialled themselves clearly in machine-readable ways. That is usually the magic circle for City of London queries, and a handful of regional firms that have done the work for regional queries. Most firms are nowhere.
What gets you cited
The pattern is consistent. Firms that get cited by ChatGPT have done three things together. They have written substantive, named-author practice area content. They have built and maintain a ReviewSolicitors presence with recent reviews. And they have deployed Person and LegalService schema so the machine-readable layer of the website matches the human-readable layer.
No one of these alone is enough. All three together produces consistent ChatGPT citation. Firms that try to game it with thin content or scraped reviews get penalised when ChatGPT is browse-the-web mode catches the inconsistency.
What this means for your firm
The honest position: if a sceptical buyer in your market asks ChatGPT for recommendations in your practice area, and your firm is not in the list, you have a visibility problem you cannot see in analytics. The query happened, the answer was given, your firm was not part of it. You only see it as a downstream absence of enquiries you should have received.
The fix is structural. Named authors, real depth, schema, ReviewSolicitors. None of it is glamorous. All of it compounds. The firms that started this work twelve months ago are now the default ChatGPT recommendation in their markets.
If you would like to see exactly how your firm appears (or does not) in ChatGPT for your priority queries, the SRA-Compliant AI Visibility Audit tests every priority query directly. Or book an intro call and we can run a few queries together.
