How Professional Services Firms Get Named by AI

What happens when a buyer asks AI to recommend a firm?

The engine builds a shortlist of firms and checks it against sources it trusts. In our test (method at the end), ChatGPT and Claude built it differently, and both read firms' own pages.

ChatGPT checks firms one by one

ChatGPT searched for firms it already knew by name plus the matter type, often only on the firm's own domain. For revenue recognition at a software company, it searched Deloitte, PwC, EY, BDO, RSM, Moss Adams and Withum this way.

Next it looked up RSM, BDO and Moss Adams in the PCAOB's register and cited the AICPA's article on revenue recognition and auditor independence.

For the Series A question it went to PitchBook's league tables of venture law firms, then searched Series A legal costs, how deferred fees work and how investor-counsel fee caps are set.

Claude reads the lists

Claude ran one broad "best firms" search per question and answered from the results. Its Series A answer cited Chambers, Vault, a third legal directory, a startup-lawyer roundup and law firms' own practice pages.

Its accounting and pricing answers cited "top firms" articles, software vendors' blogs and a few firms' guides.

What the engines did with this article's question

On this article's question, ChatGPT answered without searching. Claude cited only marketing vendors' blogs (a group that includes OnlyAEO) and no law, accounting or consulting firm.

Which pages on your own site do AI engines quote?

Pages that each answer one "firm plus matter type" search. In our pricing check, ChatGPT cited Simon-Kucher's pricing consulting page and Revenue Management Labs' pricing strategy page, each about one service.

• Practice pages by matter: A "Corporate" page can't answer "law firm for a Series A". Cooley's emerging companies page and Gunderson Dettmer's early-stage company page can, and Claude cited both.
• Guides to the problem: Claude's accounting answer cited BPM's revenue recognition guide and Burkland Associates' compliance hub. Broad searches land on guides like these.
• How an engagement runs: Phases, deliverables and staffing. We didn't test this page, but we think buyers read it before a first call.
• How you charge: ChatGPT ran fee searches (above), so explain each fee model plainly (hourly, fixed, deferred or capped) and what moves the price.

Our guide to content structure that gets cited by AI covers page layout.

How should partner and expert pages be built?

Build each around a real, named person with checkable credentials, linked from everything they write. Google strongly encourages accurate bylines where readers expect them, and asks whether they lead to background on the author and the areas they write about.

• Name, role, practice: The matter types and sectors they handle, in buyers' words.
• Verifiable credentials: Licences and designations for the profession (see the table below), named so the issuing body can confirm them.
• Representative work: Anonymized cases, written as the client-outcomes section sets out.
• Writing and speaking: Articles, talks and outside profiles that state the same facts.

Google's ProfilePage structured data documentation lists an employee page as a valid use, provided it focuses on a single person affiliated with the site. Its description holds the byline or credential, and Article markup can link to it.

Google's helpful content guidance treats made-up names and false credentials on creator profiles as deception. Anything under a partner's byline has to be work that partner stands behind.

Which outside sources decide the shortlist?

Rankings, directories, regulator registers and standards bodies decide who makes the list.

• Rankings and league tables: You can't write your way onto them, but you can make the result easy to find (see the table below).
• Regulator registers: The PCAOB's Registered Firms page finds firms by name, location or audit report activity, and each Firm Summary shows filings, inspection reports and disciplinary actions. Use the register's exact legal name on your site.
• Standards bodies: ChatGPT cited the NVCA's model legal documents as well as the AICPA. Explain your work against the standard it follows.
• Your own announcement: Cooley published its PitchBook ranking as a press release on its own news page dated 6 March 2026. Lowenstein Sandler posted PitchBook's 2025 Global League Tables ranking as firm news on 10 March 2026, and ChatGPT cited both pages.

Matching the firm's facts across these sources is entity work; see building a brand entity AI engines recognize.

Does schema markup help a law or accounting firm get cited?

Markup helps engines identify the firm, but no markup gets a firm cited on its own. Google's AI features documentation says AI Overviews and AI Mode need no special optimizations and no special schema.org structured data.

Markup does state facts without ambiguity, and Google's AI features guidance advises making sure it matches the visible text.

• Organization with sameAs: Google says Organization markup on the home or about page can help it disambiguate an organization, and sameAs points to profiles on other sites. Point it at your directory and register entries.
• The most specific type: Google recommends the most specific schema.org subtype. Law firms have LegalService (legal services, advice and representation); accounting firms have AccountingService, under FinancialService.
• Retired types: schema.org deprecated Attorney in favour of LegalService, and deprecated the general ProfessionalService type too, so check what your site template outputs.
• knowsAbout limits: schema.org says the property suggests possible expertise without implying it.

The full walkthrough is in schema markup for AEO.

How do you publish client outcomes when the work is confidential?

Publish anonymized engagement summaries in a fixed structure, and name a client only with written permission. An engine can quote a summary that gives sector, problem, work and result without naming the client.

1. Sector and client type: "A regional lender" or "a venture-backed software company".
2. The question brought: The problem in the client's terms, close to how the next buyer will phrase it.
3. What the firm did: Workstreams, team and deliverables.
4. The outcome: In terms the client would sign off on, with any figure checked against the engagement record.
5. What stays out: Any detail that lets someone in the same market identify the client.

A hypothetical example: a family-owned manufacturer asked a pricing consultancy to rebuild its spare-parts pricing after margins fell. The team reset prices by product line and retrained sales, and within the year the client's finance team reported higher realized prices.

What can law firms and financial advisers say about themselves?

Anything accurate, balanced and backed by a record, and nothing that promises an outcome; the rules below agree on that. This article isn't legal advice, so check each jurisdiction with your compliance team.

• US lawyers: The ABA's Model Rules of Professional Conduct, in the rule on communications about a lawyer's services, say a lawyer shall not make a false or misleading communication about the lawyer or the lawyer's services. Your state's version is what binds you.
• Solicitors in England and Wales: The Solicitors Regulation Authority's Transparency Rules require firms to publish price and service information for certain services, so for many firms a fees page is already an obligation.
• Investment advisers: The SEC's investment adviser marketing rule replaced the advertising rule adopted in 1961 and the cash solicitation rule adopted in 1979 with one merged rule. The SEC issued it on 22 December 2020, effective 4 May 2021.
• Broker-dealers: FINRA's rule on communications with the public requires fair and balanced communications and bars false, exaggerated or promissory claims and performance predictions.

Google gives extra weight to strong E-E-A-T on topics that could significantly affect people's financial stability, where content must be highly accurate and consistent with expert consensus. On the E-E-A-T qualities, Google's guidance says: "Of these aspects, trust is most important."

So state results with their conditions, and keep a written list of claims the firm will not make.

How do consulting, law, accounting and financial advisory firms differ?

Each profession is confirmed through a different outside record and credential, so the first page to fix differs. The financial advisory row is our own example; we didn't test that question.

Marketing agencies face their own version of this; see our guide for agencies.

How can a boutique firm compete with large firms?

By being the most specific answer for a narrow matter type, sector or jurisdiction. In our pricing check, ChatGPT searched specialists such as Holden Advisors and Wiglaf Pricing alongside McKinsey, BCG and Bain, and cited pages from Wiglaf Pricing, Blue Ridge Partners and Ducker Carlisle.

A large firm's pricing page covers every industry. A boutique can publish one on pricing for industrial distributors, by partners who have done that work, and be the closest match when a distributor asks.

How do you check whether AI names your firm?

Put your clients' questions to the engines, by practice and jurisdiction, and record who gets named and which sources are cited, rerunning the same set over time.

1. Write the buyer questions: One set per practice, sector and jurisdiction. OnlyAEO researches each buyer persona (role, company type, goals, decision criteria) because each persona gets a different answer.
2. Ask more than one engine: ChatGPT, Claude and Gemini at least, with web search on; they search differently (see the first section).
3. Log names and sources: Sort cited URLs into your site, rankings and directories, registers, and third-party lists. The empty columns show where to work.
4. Check crawler access: OpenAI's crawler documentation says sites that opt out of OAI-SearchBot won't be shown in ChatGPT search answers. OAI-SearchBot is set independently of GPTBot, and OpenAI says a robots.txt change can take about 24 hours to reach its search results. See our robots.txt guide.
5. Check Google eligibility: Google says a page must be indexed and snippet-eligible to appear as a supporting link in AI Overviews or AI Mode.

For tracking over time, see how to measure your AI visibility.

How we tested this

On 14 October 2026 we put real buyer questions to ChatGPT and Claude, with web search available, and logged their searches and citations. We run the same routine before every article: ask the buyer's question, read every cited page in full and turn each search into a section.

• Questions asked: The consulting, law and accounting questions in the table above (the accounting one also covered audit readiness), plus this article's own question.
• Limits: One run per question, US-framed questions and no Gemini. Answers change between runs, so treat these patterns as a starting point.

How long does it take, and where does OnlyAEO fit?

In OnlyAEO's client work, first measurable results usually take at least 60 days. We don't offer a results guarantee or promise specific rankings, citations or placements in AI answers.

OnlyAEO is an AEO and GEO agency. Our team works across ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Microsoft Copilot, Grok, DeepSeek, Meta AI (Muse) and Instinct: monthly AI visibility analytics, fact-checked articles on your site in your voice, help getting mentioned on the sites and listings AI reads, and a website health check.

For a regulated firm, the relevant part is the Context Graph: one record of the firm that every article is written from, including the claims it can and cannot make (the written list above). If something isn't written down, our team asks you.

The closest case to financial advisory is FastTrackr AI, a wealth management technology brand that sells software, not advice. OnlyAEO's case study page says it went from #338 to the #3 most-cited brand in wealth-tech in about 100 days. The same page shows its AI visibility rising from 0.1% to 26.7% between February and May 2026.

More questions about AI recommendations for professional firms

Should partners write their own articles?

A writer can draft around the partner's own analysis (the byline rule is in the partner pages section). Google suggests judging content by who created it, how and why, so say plainly that a named partner reviewed it.

Does the same apply to legal tech vendors?

Partly. A legal tech vendor sells software, so in-house legal teams also expect a public page on security and data handling: where data is stored, which sub-processors touch it and how breaches are reported.

Does a firm need a Wikipedia page to be named by AI?

No. None of the pages ChatGPT or Claude cited in our checks was a Wikipedia article, and Wikipedia's own notability and conflict-of-interest rules mean a firm can't simply write one for itself.

Does a Google Business Profile matter for a local practice?

For a firm that wins clients locally, such as a family law practice, yes. Google lists keeping Business Profile information up to date among the practices that stay worthwhile for its AI features.