Claude is the assistant professionals reach for when the question is considered rather than casual. In-house counsel, finance directors and company secretaries use it to shortlist external advisers, and it is embedded across enterprise tooling where those decisions get made. It searches the web and names its sources — which means your firm is either in the answer, or it is not.
Claude behaves differently from the volume engines. It is noticeably conservative about recommending a named firm, and it will hedge or decline rather than assert something it cannot corroborate. That cuts both ways. It is harder to appear at all — but once a firm is established as verifiable, it tends to be cited consistently rather than sporadically. The work is the same foundation as the rest of the AI search visibility programme, weighted towards evidence rather than volume.
Six signals that move whether Claude will name your firm.
Claude looks for the same fact in more than one place before asserting it. A claim that appears only on your own site carries little weight.
Firm name, SRA number, trading names and office addresses matching the register exactly. Mismatches make a firm look unverifiable and it gets dropped.
Substantive pages that answer the whole question outperform thin pages built around a search term. Claude summarises reasoning, so it needs reasoning to summarise.
ReviewSolicitors, Chambers and Legal 500 entries. My own 2026 research found review presence the strongest single predictor of AI citation share.
Organization, LegalService and Person schema with consistent sameAs links, so the firm resolves as one unambiguous entity.
Superlatives with nothing behind them actively hurt. “Leading” and “best” without evidence reduce confidence — which is the same standard SRA Rule 8.8 already holds you to.
Claude is the one engine where SRA compliance and AI visibility pull in the same direction. Strip out the unevidenced superlatives and you satisfy your regulator and your best-behaved AI assistant in a single edit.
Identify every claim your firm makes that exists in only one place, and build the independent sources that verify it.
Reconcile the register, your site, your directory entries and your review profiles so every source agrees.
Rewrite thin practice-area pages into pages that answer the full question, with the reasoning left in.
Build ReviewSolicitors volume and recency, and tighten Chambers and Legal 500 entries.
Remove or evidence every superlative. Improves regulatory position and citation confidence at the same time.
Priority commercial and practice-area queries tested monthly, reported alongside the other engines and your Google rankings.
Because reach is not the same as buyer quality. Claude skews heavily towards professional and enterprise use, which is where commercial instructions originate. For a firm whose work is B2B, a citation in Claude can be worth more than several in a consumer-weighted engine.
It does, but cautiously. It is more likely to describe how to choose a firm and name those it can verify than to produce a confident ranked list. That caution is precisely why verifiability work pays off here.
The foundations overlap — entity clarity, schema, SRA consistency, review presence. The weighting differs. ChatGPT rewards breadth and recognisability; Claude rewards corroboration and depth. Most of the work transfers.
Register and schema corrections register within weeks. Corroboration and review-building compound over three to six months. Depth rebuilds on priority pages typically show in months four to nine.
Yes — it sits inside the AI Search Visibility workstream of the standard SEO retainer, and it is covered in the AI Visibility Audit.
Yes. The audit tests your firm against priority queries in Claude alongside the other engines, and shows the actual answers rather than a score. If you are invisible, you will see exactly which firms are being named instead.
Each engine weights different signals. The foundations transfer between them; the emphasis does not.
Background reading: how Claude ranks UK law firms.
Book a 30-minute intro call. Or commission the SRA-Compliant AI Visibility Audit and get a formal picture across every major AI platform.
