---
title: "How professional services firms win clients from AI search"
description: "Clients of law, accounting, design and consulting firms now research with AI first. Firms get named when directories, reviews and coverage confirm expertise."
canonical: "https://underneath.agency/resources/professional-services-firms-clients-ai-search"
published: 2026-10-07
updated: 2026-10-08
publisher: "Underneath (https://underneath.agency/agent)"
entity: "https://underneath.agency/.well-known/entity.json"
---
Guide · AI search

# How do professional services firms win clients when buyers ask AI who to hire?

By making their expertise easy to confirm in the places AI assistants check: professional directories, reviews, independent coverage and clear practice pages. Across law, accounting, architecture, consulting and agencies, the same pattern is emerging. Clients use AI to understand the problem and the kind of help they need, then check names against sources they trust before they call.

## The short version

1. Legal clients start with AI: in Clio’s 2025 Legal Trends Report, as summarized by the [Illinois Supreme Court Commission on Professionalism](https://www.2civility.org/2025-clio-legal-trends-report/), 14% of consumers had used AI to answer legal questions and 43% had not but would; [Canadian Lawyer](https://www.canadianlawyermag.com/news/general/many-legal-clients-have-asked-artificial-intelligence-legal-questions-clio-report/393228) reports 28% of those who asked were told to contact a lawyer.
2. Design clients do too, then hire: in a June 2026 [Houzz survey of 3,149 renovating homeowners](https://www.constructionowners.com/press-release/houzz-survey-finds-ai-adoption-soars-among-construction-and-design-pros-while-homeowners-rely-on-the-experts), 22% had used AI for their project, and 80% still hire professionals.
3. Business buyers use AI and then verify: [Forrester](https://forrester.com/blogs/state-of-business-buying-2026) found 94% of business buyers use AI during buying, and a [Gartner survey of 645 B2B buyers](https://www.gartner.com/en/newsroom/press-releases/2026-05-20-gartner-survey-finds-sixty-nine-percent-of-b-two-b-buyers-turn-to-sales-reps-to-validate-ai-generated-insights) found 69% prefer to validate AI-generated insights with a person.
4. Assistants look up professional directories: in our study, when ChatGPT’s own search named a source such as Avvo or Super Lawyers, the answer cited that source 44.0% of the time, against 8.1% when it did not.
5. Reviews and listings shape local picks: in our local study, businesses with more reviews than the local median were 19.5 points more likely to be listed by ChatGPT, and accountants and lawyers were among the professions tested.

This is the overview. For one profession in depth, read our articles on [accounting firms](https://underneath.agency/resources/accounting-firms-clients-ai-search), [IT services firms](https://underneath.agency/resources/it-service-firms-leads-ai-search), [managed service providers](https://underneath.agency/resources/msps-customers-from-ai-search), [penetration testing firms](https://underneath.agency/resources/pentest-firms-leads-ai-search) and [wealth management firms](https://underneath.agency/resources/wealth-management-firms-clients-ai-search). Nothing here is legal, tax or financial advice.

## Why is buying professional services different from buying products?

Because the client is buying judgment they cannot inspect in advance, so trust and proof come before price.

A buyer can test software in a trial. They cannot test a lawyer, an auditor or an architect until the work is underway. That is why referrals still dominate. In a 2025 TaxDome survey of 350 US businesses reported by [CPA Practice Advisor](https://www.cpapracticeadvisor.com/2025/08/19/survey-of-smbs-shows-how-they-choose-and-evaluate-their-accountant-firm/167532/), 57% found their current accountant through a peer referral and only 3% through advertising.

Professional services also have two kinds of buyer, and AI sits differently with each:

- **Individuals and owners** hiring a lawyer, tax preparer, architect or advisor for a personal or small-business matter. They often do not know what kind of professional they need, so they ask.
- **Business buyers**, such as a general counsel, CFO, CMO or COO, hiring a firm through a shortlist, pitch or request for proposal. Forrester reports that, on average, 13 internal stakeholders and nine external participants influence business buying decisions.

In both cases, a referral now tends to be checked rather than simply accepted. We infer that AI assistants have joined the places that check happens, alongside the firm’s website, reviews and peers.

## Where does AI already sit in how clients choose a firm?

At the start, where clients work out the problem and the help they need. Increasingly, also at the check.

The evidence differs by profession, but it points the same way:

| Profession | What the evidence shows | Where to read more |
|---|---|---|
| Law | 14% of consumers have used AI for legal questions; 28% of those who asked were told to contact a lawyer (Clio, 2025) | This article |
| Accounting and tax | Tax-related ChatGPT searches in early 2026 were four times the year before; 30 percent were for help filing forms and using tax software ([GovTech](https://www.govtech.com/question-of-the-day/how-often-did-americans-turn-to-ai-for-help-with-their-taxes-this-year)) | [Accounting firms](https://underneath.agency/resources/accounting-firms-clients-ai-search) |
| Architecture and design | 22% of renovating homeowners used AI; comparing options was a use for 50% of them (Houzz, 2026) | This article |
| Consulting and advisory | 30% of small business owners go to generative AI such as ChatGPT for advice ([TD Bank and Wakefield Research](https://stories.td.com/volumes/default/Wakefield-Research-Analysis-of-Results-for-TD-Bank-4.10.25.pdf), 2025) | This article; [business advisory firms](https://underneath.agency/resources/business-advisory-firms-clients-ai-search) |
| Agencies and IT services | 45% of B2B buyers used generative AI, mainly to gather information on vendors and products (Gartner, 2026) | [IT services](https://underneath.agency/resources/it-service-firms-leads-ai-search), [MSPs](https://underneath.agency/resources/msps-customers-from-ai-search) |

Two patterns cut across the table. First, AI shows up early: clients use it to imagine, understand and compare. Houzz found homeowners used AI for renovation or design ideas (64%), visualizing concepts (58%) and comparing options (50%), far more than for budgeting (24%) or contracts (6%). Second, clients still hire people. Among homeowners who chose not to use AI at all, the top reason was a preference for professional expertise (38%).

## What do clients ask AI before they contact a firm?

Four kinds: the problem, who handles it, how the options compare, and whether a firm is any good.

We wrote the examples below to show typical phrasing; none comes from a real client’s chat.

| Profession | Illustrative prompt |
|---|---|
| Law | “My business partner is locking me out of the company accounts. What kind of lawyer do I need in Austin?” |
| Accounting | “Regional CPA firm vs Big Four for a first SOC 2 report at a 60-person SaaS company” |
| Architecture | “Architects in Denver who specialize in mid-century home remodels, and what they typically charge” |
| Consulting | “Who helps mid-sized distributors fix inventory planning, and how do I compare them?” |
| Agency | “B2B content agencies for cybersecurity companies with good client reviews” |
| Any profession | “Is [firm name] reputable? What do former clients say?” |

The last question matters most for firms that already win on referral. A prospect who hears a firm’s name from a colleague can now ask an assistant about it in seconds, and the answer draws on whatever the web says about the firm.

## How does an AI answer turn into an inquiry or a shortlist?

Through a short chain: the answer frames the problem, names options, and the client verifies a name before reaching out.

For individuals, the path tends to run from a problem question → an answer that explains what kind of professional to hire → a follow-up asking who → a firm page or directory profile → a consultation request. The Clio finding that 28% of people who asked AI legal questions were told to contact a lawyer shows the hand-off happening.

For business buyers, the path runs from research → a shortlist → a pitch or proposal → a retainer or project. Here the person matters at the end. Gartner found buyers used an average of seven information sources during a recent purchase, and 69% prefer to validate AI-generated insights with a sales rep. In a professional firm, that “rep” is usually a partner or principal. Forrester adds that buyers are more likely to engage with providers based on information from industry experts rather than AI tools.

Trust cuts both ways. Gartner found 51% of buyers say they are more likely to encounter misleading information from generative AI, and 49% say the same of a sales rep. Our reading: AI can put a firm on the list, but proof from independent sources and people is what keeps it there.

## What decides which firms an AI assistant names?

The sources it searches: professional directories, rankings, reviews, independent coverage and clear practice pages.

Here is what is documented by the platforms, observed in studies, and inferred by us.

- **Documented by Google.** [Google’s guidance on AI features](https://developers.google.com/search/docs/appearance/ai-features) says AI Overviews and AI Mode may use “query fan-out,” issuing multiple related searches across subtopics and data sources to develop a response.
- **Documented by OpenAI.** [ChatGPT search](https://help.openai.com/en/articles/9237897-chatgpt-search) typically rewrites a question into one or more targeted queries for its search partners, and may share general location to make results local.
- **Observed in our hidden-searches study.** Asked “Who is the best divorce lawyer in Houston?”, ChatGPT ran four searches, naming Avvo, Martindale-Hubbell, U.S. News Best Lawyers and Super Lawyers. [Across the study](https://underneath.agency/research/ai-hidden-searches-study), when a search named a source, the answer cited it 44.0% of the time.
- **Observed in our local study.** ChatGPT listed [67.7% of the businesses ranked 1 to 3 on Google Maps](https://underneath.agency/research/chatgpt-local-picks-google-profile-study) for local questions about five services, including accountants and personal injury lawyers, and review volume was the one signal that held everywhere.
- **Observed in our brand studies.** Each tenfold increase in independent sites naming a brand went with [4.7 times the odds of being recommended](https://underneath.agency/research/brand-entity-ai-recommendations-study), and 88.0% of answers to “is this brand legit?” [cited a review or complaint platform](https://underneath.agency/research/is-it-legit-ai-reputation-study).

Each profession has its own version of those sources. Law has legal directories and peer rankings. Accounting has state society and CPA listings. Architects and designers have portfolio platforms such as Houzz. Agencies and development firms have review marketplaces such as Clutch, which in 2026 launched an app inside ChatGPT that brings [verified provider reviews into AI conversations](https://www.demandgenreport.com/?p=52857). We infer that a firm’s profile on its profession’s main directories now does double duty: it is read by the client and by the assistant.

## What happens when an assistant leaves out or garbles a firm’s details?

Missed consultations and shortlist places, plus wrong details that misdirect clients. Nobody has published the revenue cost yet.

**Being misdescribed.** When we asked four AI engines for the address, phone, website and hours of real local businesses, answers about accountants [differed from the Google profile 17.7% of the time](https://underneath.agency/research/ai-business-facts-accuracy-study), and answers about personal injury lawyers 16.4%. A wrong phone number on an intake call is a lost client.

**Being inconsistent.** In [our consistency study](https://underneath.agency/research/ai-recommendation-consistency-study), only 25.2% of the brands ChatGPT named for a question appeared in all five runs. A firm checked once may look fine and then disappear on the client’s own search.

**Being absent.** The cost of not being named is our inference, not a measured loss: the client who learns the options from an assistant meets the named firms first. For a business buyer, that can mean the shortlist forms before a partner ever hears about the opportunity. Our article on [enterprise software shortlists in AI search](https://underneath.agency/resources/enterprise-software-shortlists-ai-search) shows the same thing happening in a neighboring market.

## How does GEO work for a professional services firm?

Generative engine optimization (GEO) makes a firm’s expertise, people and proof easy for AI assistants to find, confirm and repeat.

1. **Problem-first practice pages.** One page per problem the firm solves, in the client’s words, naming the industries, places and kinds of matter or project it takes on.
2. **The profession’s directories.** Complete, consistent profiles on the directories that matter in your field: legal directories, CPA and society listings, Houzz for design, Clutch for agencies and development firms.
3. **People pages with credentials.** Bar admissions, licenses, certifications, publications and industries served, matched across the firm’s site, LinkedIn and directory profiles.
4. **Independent coverage.** Trade press bylines, association talks, rankings and awards from credible publishers. This is where [authority for AI search](https://underneath.agency/resources/how-brands-build-authority-for-ai-search) comes from.
5. **Reviews and client proof within professional rules.** Reviews where the profession allows them, case studies or portfolios with client permission, and anonymized outcomes where it does not.
6. **Consistent facts.** The same name, offices, phone and practice list everywhere. If AI answers already get something wrong, [here is how to correct it](https://underneath.agency/resources/fix-wrong-brand-information-in-ai-answers).
7. **Regular testing.** Ask the problem, who, compare and check questions across ChatGPT, Gemini, Perplexity, Claude and Google AI Mode, several times each, every quarter.

Clients increasingly care how firms use AI, too. In Clio’s report, 78% of clients want lawyers to disclose AI use. Lawyers, accountants and registered advisers must also keep every public claim within their profession’s advertising rules. Those rules are no handicap here, because sober facts anyone can check are exactly what an assistant can pass on without distortion. No firm can be promised a place in an AI answer; what GEO does is make the record about its lawyers, accountants or designers accurate, specific and easy to locate.

## What is still unknown about AI search and professional firms?

Mostly the last step: how often an AI mention becomes a signed engagement, and how much each profession differs.

- **No engagement data.** We found no public data linking a firm’s AI visibility to consultations, proposals or signed work, in any profession.
- **Vendor surveys.** The client-behavior figures come largely from software companies and platforms with an interest in the topic, such as Clio, Houzz and TaxDome.
- **Uneven coverage.** Law and accounting have the most evidence; architecture, consulting and agencies have little that is independent and public.
- **Firms are changing too.** The [Thomson Reuters Institute](https://www.thomsonreuters.com/en-us/posts/technology/ai-in-professional-services-report-2026) found organization-wide AI use in professional services almost doubled to 40% in 2026, so what clients expect from firms is shifting at the same time as how they find them.

## Where should a professional services firm start?

Pick the three practice areas or services you most want to grow. Then ask AI assistants what a client would ask.

For each, write the problem question, the who question, the comparison question and the reputation question, including the place where it matters. Run them in ChatGPT, Gemini, Perplexity and Google AI Mode several times. Note which firms are named, which directories and pages are cited, and whether your firm’s facts are right. That gap is the plan.

If your partners want more consultations, shortlist places and proposal invitations from clients who now research with AI first, [ask us for an AI visibility review of your firm](https://underneath.agency/contact). We will test the questions your clients ask, show the directories and sources assistants rely on in your profession, and plan the pages, profiles and coverage that connect your expertise to those clients. Our [generative engine optimization service](https://underneath.agency/services/generative-engine-optimization) page explains how that work runs for a firm, from practice pages and directory profiles to quarterly testing of client questions.

## Frequently asked questions

### Do people use ChatGPT to find a lawyer or an accountant?

Increasingly. 14% of consumers have used AI for legal questions, and in our study ChatGPT answered “best accountant” and “best lawyer” questions with named local firms, addresses and ratings.

### Which directories matter most for AI visibility?

It depends on the profession. In our study ChatGPT searched Avvo, Martindale-Hubbell and Super Lawyers for a lawyer question; agencies have Clutch, designers have Houzz.

### Do referrals still matter if clients use AI?

Yes. 57% of businesses found their accountant through a referral. AI is increasingly where a referred name gets checked, so the evidence about the firm must hold up.

### Should a firm publish content written for AI?

Write for clients. Clear pages about the problems you solve, who you serve and who does the work help clients and assistants alike; no format guarantees a mention.

## Sources

- Illinois Supreme Court Commission on Professionalism (2025), [Future Law: 2025 Clio Legal Trends Report](https://www.2civility.org/2025-clio-legal-trends-report/)
- Canadian Lawyer (2025-10-17), [Many legal clients have asked artificial intelligence legal questions: Clio report](https://www.canadianlawyermag.com/news/general/many-legal-clients-have-asked-artificial-intelligence-legal-questions-clio-report/393228)
- Houzz via Construction Owners (2026-09), [Houzz Survey Finds AI Adoption Soars Among Construction and Design Pros, While Homeowners Rely on the Experts](https://www.constructionowners.com/press-release/houzz-survey-finds-ai-adoption-soars-among-construction-and-design-pros-while-homeowners-rely-on-the-experts)
- Forrester (2026), [The State Of Business Buying, 2026](https://forrester.com/blogs/state-of-business-buying-2026)
- Gartner (2026-05-20), [Gartner Survey Finds 69% of B2B Buyers Turn to Sales Reps to Validate AI-Generated Insights](https://www.gartner.com/en/newsroom/press-releases/2026-05-20-gartner-survey-finds-sixty-nine-percent-of-b-two-b-buyers-turn-to-sales-reps-to-validate-ai-generated-insights)
- CPA Practice Advisor (2025-08-19), [Survey of SMBs Shows How They Choose and Evaluate Their Accounting Firm](https://www.cpapracticeadvisor.com/2025/08/19/survey-of-smbs-shows-how-they-choose-and-evaluate-their-accountant-firm/167532/)
- GovTech (2026-04-15), [How often did Americans turn to AI for help with their taxes this year?](https://www.govtech.com/question-of-the-day/how-often-did-americans-turn-to-ai-for-help-with-their-taxes-this-year)
- TD Bank and Wakefield Research (2025-04), [TD Bank Financial Preparedness Survey: analysis of results](https://stories.td.com/volumes/default/Wakefield-Research-Analysis-of-Results-for-TD-Bank-4.10.25.pdf)
- Demand Gen Report (2026-05-12), [Clutch Launches First B2B Services Marketplace App on ChatGPT](https://www.demandgenreport.com/?p=52857)
- Thomson Reuters Institute (2026), [AI in Professional Services Report 2026](https://www.thomsonreuters.com/en-us/posts/technology/ai-in-professional-services-report-2026)
- Google Search Central (2025), [AI features and your website](https://developers.google.com/search/docs/appearance/ai-features)
- OpenAI Help Center (2026), [ChatGPT search](https://help.openai.com/en/articles/9237897-chatgpt-search)
- Underneath (2026), [The hidden searches AI assistants run before they answer](https://underneath.agency/research/ai-hidden-searches-study)
- Underneath (2026), [Which Google Maps businesses does ChatGPT recommend?](https://underneath.agency/research/chatgpt-local-picks-google-profile-study)
- Underneath (2026), [Do AI answers match a business’s Google profile?](https://underneath.agency/research/ai-business-facts-accuracy-study)
- Underneath (2026), [Do Wikipedia and schema make AI assistants recommend a brand?](https://underneath.agency/research/brand-entity-ai-recommendations-study)
- Underneath (2026), [“Is this brand legit?” How AI assistants build a reputation](https://underneath.agency/research/is-it-legit-ai-reputation-study)
- Underneath (2026), [Ask an AI the same question 5 times: do the brands change?](https://underneath.agency/research/ai-recommendation-consistency-study)

---

This is the Markdown twin of https://underneath.agency/resources/professional-services-firms-clients-ai-search. The HTML page is canonical. Publisher: Underneath, https://underneath.agency/agent. Site index: https://underneath.agency/llms.txt.