Professional-services firms can appear in AI answers through their own service pages as well as directories, reviews and third-party coverage. Start with a specific client need, geography or specialty, then inspect which firms and sources appear. The evidence for that question should guide the work; off-site coverage is not the only possible source.
Keep both owned and independent pages in the plan. Your service pages can explain capabilities, constraints, pricing and evidence of experience. Reviews and independent coverage can add other perspectives. Check the cited pages for your actual buyer questions before deciding which source deserves attention.
Why professional services firms go invisible in AI search
A search ranking and an AI recommendation are different observations. Muck Rack’s May 2026 study analyzed more than 25 million links across ChatGPT, Claude and Gemini in 17 industries. It classified 84% of citations as earned media and 0.3% as paid or advertorial content. Its earned-media category includes journalism, academic research, government sources, encyclopedic sites and third-party corporate content; the figure is not an 84% share for press coverage alone.
That broad study does not establish the source mix for professional-services recommendations. A consultancy’s service pages and biographies can explain its expertise, while client reviews and independent reporting can supply additional evidence. If a firm is missing from a particular answer, inspect the question, named firms and cited pages before attributing the gap to its website or PR activity.
Where AI looks when a buyer asks for a firm
Source mixes differ by platform. Profound’s analysis of 680 million citations from August 2024 to June 2025 compared ChatGPT, Google AI Overviews and Perplexity. Among each platform’s top ten sources, Gartner represented 7.0% of Perplexity’s citations and LinkedIn 13.0% of Google AI Overviews’ citations. Those are shares within the top-ten lists, not shares of every citation or a professional-services-only sample.
Alongside your own service pages, these three off-site source types are worth checking. The table gives examples and maintenance tasks, not a measured ranking of their effect.
| Source type | Source examples | Maintenance task |
|---|---|---|
| Service directories | Clutch, Gartner Peer Insights, G2 profiles with reviews and category data | Incorrect or incomplete profile facts |
| Earned media | Named-partner commentary in trade and business press | Check named expertise and factual claims |
| Entity sources | LinkedIn, Crunchbase, Wikipedia, consistent firm and partner naming | Inconsistent names, titles, and descriptions across the web |
Use the table to investigate your own question set. A cited directory profile may need a factual correction; a cited article may describe a competitor’s specialty more clearly than yours. Neither observation proves that buying a listing or earning coverage will change a recommendation. Our research on who publishes the pages AI cites compares publisher types with explicit study boundaries.
How buyers now build the shortlist
G2’s March 2026 survey of 1,076 B2B software buyers provides context for how buyers use AI. In that survey, 71% used chatbots for software research and 51% began with a chatbot more often than Google, up from 29% in April 2025. G2 reported 63% for ChatGPT in software research. Some 69% said AI guidance led them to choose a different vendor, and 33% bought from a company they had not previously heard of. These are self-reported software-buyer results, not measured outcomes for professional-services firms.
B2B software buyers researching vendors with AI
Software buyers who changed their planned vendor choice
Citations in Muck Rack’s earned-media category
Software buyers most reassured by review-site citations
In the same software-buyer survey, 45% identified software-review-site citations as the most confidence-inspiring signal in an AI response. That is a reason to examine relevant reviews, not evidence that directories are the only useful source for service firms.
If you want to see which sources shape AI answers about your brand, run a free brand check. It takes a minute.
Start with the prompts that map to your practice areas
Before optimizing anything, define the questions you need to win. A professional services prompt set is organized by practice area and buyer situation, not by keyword volume. For each service line, write the open-ended questions a buyer would actually ask: "best cybersecurity consultancy for healthcare," "top employer brand agencies for hiring at scale," "who can audit our Series B financial model."
Aim for 20 to 40 prompts that map directly to revenue. Include unbranded category questions (where you are invisible today), comparison questions ("X firm vs Y firm"), and situation questions tied to a trigger event like a funding round or a compliance deadline. This set becomes the measurement baseline: you run it across models, record which firms get named and which sources get cited, and you re-run it on a fixed cadence. Without a defined prompt set, "are we visible in AI" stays an anecdote. With one, it becomes a number you can move. Our guide on building an AI visibility prompt set covers the selection mechanics in depth.
Check the directories cited for your services
Directory profiles can combine reviews, category information and service details in one readable place. G2’s acquisition of Capterra, Software Advice and GetApp in early 2026 and Clutch’s April 2026 ChatGPT marketplace app show that these services are evolving. Clutch’s app uses its marketplace data; appearing there is a separate outcome from being recommended in an ordinary ChatGPT answer. Assess the exact surface before spending on it.
Check whether a directory appears in answers to your target questions and whether its category fits your firm. Keep relevant profiles accurate and ask clients for honest accounts of their work with you. Track citations and recommendations separately; more reviews or a paid listing do not guarantee either.
Turn partner expertise into earned-media citations
Independent coverage can make a partner’s expertise easier to examine. Journalism accounted for 27% of citations in Muck Rack’s May 2026 study. That finding supports checking relevant publications, but it does not measure the effect of a new placement on your firm’s recommendations.
Offer specific, attributable commentary that a publication’s readers can use. Choose outlets relevant to your practice and lead with claims you can substantiate. Track whether published coverage is cited and whether your firm is named in later answers. If you want help with this work, Soar offers managed AI visibility. Our earned media for AI citation guide covers the approach.
Build the entity layer so AI knows who you are
Keep the firm’s name, website, partner titles and service descriptions consistent across pages you control. Link partner biographies to supported profiles and publications. Clear company facts help readers distinguish similarly named firms; check AI answers for identity errors rather than assuming that markup or a directory entry resolved them.
Standardize the firm name, partner names, titles, and a one-line description, then apply them everywhere. Add Organization and Person schema with sameAs links to your authoritative profiles so machines can connect the dots. Consistent names help search and AI systems connect your listings, coverage and reviews to the same firm. Our guide to entity disambiguation for AI covers the schema and naming mechanics.
A placement may be cited quickly, later, or never. Repeat comparable questions and inspect the answers over time. Firms in regulated practices should review their published service and outcome claims through their existing compliance process. See our playbook for regulated industries.
Measure citation gaps, not just rankings
Knowing earned media matters is not the same as knowing which sources name your competitors and not you. A citation gap analysis answers that: run your prompt set across models, capture every source the answer cites, and compare the sources backing competitors against the sources backing you.
Use repeated source gaps to prioritize investigation. If competitors appear in a cited Clutch category page or industry article and your firm does not, assess whether you belong there and whether the page is accurate. A citation gap is an observed difference, not proof of why a competitor won or what will move pipeline. Our AI citation gap analysis guide explains how to make that comparison.
Run it on an operating cadence
AI recommendations shift continuously, so a one-time audit decays. Treat AI visibility as a recurring marketing function with a fixed rhythm. Monthly, re-run the prompt set, record which firms and sources appear, and check directory review velocity. Quarterly, refresh the earned-media target list against the citation gaps, audit entity consistency for any new partners or office changes, and report movement to firm leadership in the same review where you present pipeline.
Give the marketing lead responsibility for the question set and review, with practice partners supplying expertise and PR helping with relevant coverage. Record the changes you make, retain the answers and sources, and compare like periods. This makes it easier to distinguish a repeated pattern from an isolated answer.
Who should use this playbook
This playbook applies to firms selling expertise, including consultancies, agencies, accounting practices, IT service providers and staffing firms. If relevant buyer questions omit your firm, use the answers to investigate its service pages, identity and independent coverage together.
Adapt the source choices to the practice. A solo specialist may have useful publications but no relevant directory profile; a regulated firm needs its existing claims review. Keep the sequence simple: define buyer questions, inspect the evidence, fix supported gaps and measure again. None of these steps guarantees a recommendation.