By Steve Merrill, Founder of WRKNG Digital — November 2, 2026
How does AI decide which consulting firm to recommend?
AI assistants recommend the consulting firm they can verify and cite. When a buyer asks ChatGPT or Perplexity "who should I hire for X," the model picks firms with third-party mentions on trusted sites, consistent business data, strong reviews, and specific service content it can quote. It skips the ones it can't confirm.
Professional services have always run on trust and referrals. A friend names a firm. A partner at another company vouches for them. That's how the work moved.
AI is now a referral engine. And it works on the same instinct a person does. It wants proof before it puts its name on a recommendation.
Most firms fail that test. Not because they're bad at the work. Because their website says "we deliver excellence" and nothing a machine can actually repeat.
Why do third-party mentions matter most?
Third-party mentions are the strongest signal because they're proof you didn't write about yourself. When an authoritative site names your firm, AI reads it as a vote. The more trusted sites that mention you in the right context, the safer you look to recommend.
Think about how you'd answer if a friend asked you to name a good tax firm. You wouldn't quote the firm's own homepage. You'd say who you've heard other people trust.
AI does the same thing. It weighs what the open web says about you far more than what you say about yourself.
This is where firms get stuck. You can control your website all day. You can't control whether Forbes, a trade publication, or your industry association ever names you. You have to earn that.
The firms winning in AI answers are the ones other sites already talk about.
Google's own guidance has pushed this for years through what they call experience, expertise, authoritativeness, and trust. You can read it in their helpful content documentation. The models absorbed that same logic.
What does consistent entity data actually mean?
Entity data is your firm's name, address, and phone number, plus how you describe what you do. When those match across your site, Google Business Profile, LinkedIn, and directories, AI reads you as one real business. When they conflict, it hesitates.
I've seen this kill good firms. One office lists the suite number, another drops it. The LinkedIn page uses a different legal name than the website. A directory still has the old phone.
Small stuff. Except the machine can't tell if that's one firm or three. So it stays vague, or it names a competitor it's sure about.
Pick one exact version of your name, address, and phone. Make every listing match it. This is the cheapest fix on this list and almost nobody does it.
How do reviews and reputation feed the answer?
Reviews tell AI you've done the work and people were happy. Volume, recency, and the specific words clients use all get read. A firm with recent, detailed reviews that mention real services looks more trustworthy than one with five old stars and no text.
Here's the part firms miss. The language inside a review is data. When a client writes "they rebuilt our forecasting model and we closed the quarter early," that sentence tells AI exactly what you do and that it worked.
Ask clients for reviews that name the service and the result. Not "great to work with." That tells the machine nothing.
Why does vague content lose to specific content?
AI can only cite what it can quote. A page that says "we deliver strategic excellence" gives it nothing to repeat. A page that says "we cut a manufacturing client's close time from 12 days to 4" gives it a sentence it can use in an answer.
This is the biggest self-inflicted wound in professional services. Every firm's site reads the same. Trusted advisor. Tailored solutions. Deep expertise. Blank.
When a buyer asks "best consulting firm for supply chain cost reduction," the model needs a firm that has published about supply chain cost reduction, with a client type and a number. If your pages are abstract, you're invisible for the exact query you want.
Write pages around a service plus an outcome. Name the industry. Name the problem. Name what changed. That's what gets pulled into an answer.
What role does structured data play?
Structured data is code that labels your pages so machines read them cleanly. Organization schema confirms who you are. Service schema spells out what you offer. FAQ schema feeds direct answers. It removes guesswork for the AI that's deciding whether to trust you.
You can see the exact formats at schema.org. It's free and it's the shared vocabulary every major assistant understands.
Schema won't save weak content. A firm with vague pages and perfect markup still has nothing worth quoting. But strong content with clean schema is far easier for a model to parse, trust, and cite.
Microsoft's Bing team has said the same thing about markup helping engines understand pages, which you can find in their structured data guidance.
Does recency change who gets recommended?
Yes. AI leans toward current signals. A firm with reviews from last month, a recently updated services page, and fresh third-party mentions looks active and safe. Stale pages and old reviews read as a firm that might not even be operating the same way.
You don't need to publish daily. You need to not look frozen. Update your core pages. Keep reviews coming in. Refresh case numbers when they change.
How can a firm actually fix this?
Start with what you control, then earn what you don't. Here's the order that works.
- Fix your entity data. One exact name, address, and phone everywhere.
- Rewrite your service pages around outcomes. Industry, problem, result, number.
- Add Organization, Service, and FAQ schema. Clean markup on every core page.
- Build a review engine. Ask for reviews that name the service and the result.
- Earn third-party mentions. Press, trade publications, associations, guest articles in your niche.
- Keep it fresh. Update pages and keep new reviews flowing.
Do the first four this quarter. They're inside your walls. The fifth is the long game, and it's the one that moves you from "a firm" to "the firm AI names."
Where does this leave professional services?
The referral never went away. It just added a new source. Your next client might hear your name from a partner at dinner, or from ChatGPT at 11pm while they're building a shortlist.
Both run on the same thing. Proof you can be trusted. The firms that publish specific, verifiable content and earn real mentions will get named. The ones hiding behind "we deliver excellence" will keep wondering why the phone stopped ringing.
Trust was always the product. Now a machine reads it too.
Frequently asked questions
Does AI recommend consulting firms based on who pays the most?
No. Assistants like ChatGPT, Perplexity, and Gemini pull from public text and citations. They recommend firms mentioned on trusted third-party sites, with consistent entity data and specific content about what they do.
Can a small firm get recommended over a big one?
Yes. A small firm that owns a narrow specialty, earns mentions in that niche, and publishes clear outcome content can get cited ahead of a large generalist for specific queries.
How long does it take to show up in AI answers?
Entity and schema fixes can get picked up within weeks. Third-party mentions and reputation take months, because they depend on other sites publishing about you.
What matters more, my website or outside mentions?
Both, and they work together. Your site proves what you do and feeds schema. Outside mentions prove you're trusted. AI needs the second to feel safe recommending you.
Want to become the firm AI recommends?
We help professional services firms turn their site and reputation into signals AI can actually cite. See how it works at WRKNG Digital.
