By Steve Merrill · October 30, 2026
Why don't accounting and consulting firms show up when people ask AI for recommendations?
Because your website gives the model nothing to work with. Thin service pages, no direct answers it can lift, weak local and entity signals, and almost no presence on the sites AI already trusts. When a prospect asks ChatGPT or Perplexity for a good firm, the model picks whoever gave it the cleanest, most specific answer. That's rarely you.
I'll be honest. When I first dug into this for service firms, I assumed it worked like Google. Rank the page, win the client. It doesn't. An AI assistant doesn't show ten results. It reads a few sources and writes one answer that names two or three firms. If you're not in that answer, you don't exist to that buyer.
Here's the bottom line: your competitors aren't beating you on reputation. They're beating you on readability.
What makes a firm invisible to AI in the first place?
Four things, almost every time. Most professional services sites are built to impress a human skimming for thirty seconds, not a machine trying to extract a clear answer.
Here's what I see in audit after audit.
- Thin service pages. "We offer tax, audit, and advisory services for businesses of all sizes." That sentence tells a model nothing. No industries, no problems, no specifics.
- No structured answers. The page never actually answers a question a buyer would type. It describes the firm. It doesn't reply to anything.
- Weak local and entity signals. Mismatched addresses, a half-finished Google Business Profile, no clear statement of who you are and where you work.
- No third-party presence. Nothing on the directories, review sites, or trade publications the models pull from. The only place you describe yourself is your own site.
Put those together and the model has no reason to trust you and nothing clean to quote. So it names the firm that gave it both.
How big is the shift to AI search really?
Big enough that you can't wait this out. Google confirmed its AI Overviews reached over a billion users in 2024, and the research step for hiring a professional is moving into the assistant itself.
Gartner has predicted traditional search volume will drop 25% by 2028 as buyers shift to AI chatbots. Perplexity built its whole product around cited answers, and ChatGPT now searches the web mid-conversation. When a business owner asks "who's a good CPA for a SaaS company in Denver," a model is writing that answer right now. The only question is whose name is in it.
Step 1: Rewrite your service pages as actual answers
Your service pages have to answer the real questions a buyer asks, with the direct answer in the first two sentences. Lead with the reply. Then prove it. A model can lift a clear answer. It can't lift a brochure.
Take a vague page like "Audit Services" and turn it into something a prospect would search. "How does a SaaS company prepare for its first financial audit?" Answer it in plain language. Name the industries you serve and the specific problems. The more concrete the page, the more a model has to grab.
I worked with a consulting firm that had one page covering everything they did. We split it into five, each answering one narrow question their best clients had asked on sales calls. Specificity did the work.
Step 2: Add the structured markup a model can read
Schema markup tells a machine what you do, where, and for whom, in a format it doesn't have to guess at. Add Organization, Service, FAQPage, and local business schema to your site. It makes your answers explicit instead of implied.
Most firm sites have zero structured data. The page looks fine to a person and reads as a blank to a model. You can check what Google and the assistants see with the Schema.org validator. Fix what's missing, then add FAQ schema to every service page so your answers are machine-readable.
This part isn't glamorous. It's plumbing. But plumbing is why one firm gets quoted and the firm next door gets skipped.
Step 3: Fix your entity and local signals
Entity signals are how a model confirms who you are, and local signals confirm where you work. Clean up your Google Business Profile, your directory listings, and your name-address-phone consistency across the web. Mismatched details make a model unsure, and an unsure model picks someone else.
For a lot of firms, this is the fastest win. One firm I looked at had three different addresses floating across old directory listings. We standardized every one. Their visibility in local AI answers climbed within weeks, before we touched a single blog post.
Make your name, address, phone, and service area identical everywhere. Fill out the Google Business Profile completely. Boring work. It matters more than most of the shiny stuff.
Step 4: Get present on the sources AI already trusts
AI assistants don't only read your website. They pull from directories, review platforms, and trade publications they already weight heavily. For professional services that means your Google Business Profile, Clutch, industry association listings, and the publications your buyers read.
When a prospect asks for a recommendation, the model often leans on review sites and directory roundups as much as any firm's own page. So earn the mentions. Get real client reviews. Claim every relevant listing. If you've got a specialist angle, pitch it to a trade publication.
Your own site tells the model what you say about yourself. These sources tell it what the market says about you. The second one carries more weight.
Step 5: Publish content that reads like a specialist
Models favor specialists over generalists. Publish deep, specific content on the narrow problems your best clients hire you to solve, so a model reads you as the firm that owns that problem. Depth signals expertise. A thin "about our services" page signals nothing.
Pick the three problems you solve better than anyone. Write the definitive piece on each. "How should a dental practice structure its books for a sale?" "What does a Series A company need from a fractional CFO?" Answer the question your ideal client is typing into an assistant.
Don't try to cover everything. A firm known for one thing gets named for that thing. A firm that does everything gets named for nothing.
How do you know it's working?
Run your own test. Once a week, ask ChatGPT, Gemini, and Perplexity the exact questions a prospect would ask to find a firm like yours. Log whether you get mentioned, and whether you get cited with a link.
Mentions come first. Citations follow. When your name starts appearing in the answer, and then your link shows up under it, the work is landing. That's your signal to go deeper on whatever topic is winning.
No agency or tool saves you here. You can have every fix on this list in front of you and still never rewrite the one service page that would've gotten you quoted. Pick up the tool and turn the bolt.
Frequently asked questions
Why don't accounting and consulting firms show up in AI answers?
Most firm websites are thin on specifics, give no direct answers a model can lift, carry weak local and entity signals, and have almost no presence on the third-party sources AI assistants trust. The model has nothing clean to read, so it names someone else.
Does a professional services firm need a blog to get cited by AI?
Not strictly, but specialist content helps a lot. Models favor firms that cover a narrow problem deeply over firms with one vague services page. A handful of detailed pages on the exact problems you solve usually does more than a general blog.
How long before a firm starts showing up in ChatGPT or Perplexity?
In the firms I've worked with, cleaning up entity and local signals plus rewriting service pages tends to move the needle in 60 to 120 days. Third-party mentions and reviews often show up faster because the models already weight those sources.
Do reviews on Google and directories affect AI recommendations?
Yes. AI assistants lean on Google Business Profiles, directories, and review platforms when a prospect asks for a local or specialist firm. Consistent listings and real reviews give the model a reason to name you.
Want help becoming the firm AI recommends?
This is the exact work we do at WRKNG Digital. If you want your firm built to show up when prospects ask ChatGPT, Gemini, and Perplexity for a recommendation, start here: wrkngdigital.com/agentic-commerce-landing-page.
