For years, tax and accounting firms have measured online visibility in familiar ways. Where do we rank on Google? Are we showing up in Maps? Is organic traffic growing? Are our reviews strong?
Those questions still matter. But they no longer tell you everything about how a prospective client may discover your firm.
There is another question worth asking now:
When someone asks AI who they should hire, does AI know enough about your firm to recommend you?
That is becoming a measurable marketing question, not a theoretical one.
Your next client may not search the way your last client did
Imagine a business owner sitting at home after dinner.
A few years ago, that person might have opened Google and typed:
CPA near me
Today, they can open ChatGPT, Gemini, Perplexity or Google's AI-powered search and explain the real problem:
I own a construction company doing about $4 million a year. My accountant mostly prepares my return, but I need someone who understands proactive tax planning, succession and eventually selling the company. How do I find the right CPA?
That interaction contains far more information than a traditional keyword search. The prospect has described the business, the problem, the stage of life and the kind of advisor they want.
AI can then help research the issue, identify firms, compare them and narrow the choices before the prospect ever contacts anyone.
The broader data suggests this behavior is already moving into the mainstream. McKinsey research on AI-powered search found that 50% of U.S. consumers intentionally use AI-powered search and projects that $750 billion in U.S. consumer spending could flow through AI-powered search by 2028. That's consumers broadly, not CPA buyers, but the direction is clear.
The B2B data gets even closer to the way professional services are purchased.
Gartner's May 2026 B2B buyer research surveyed 645 buyers and found that 45% had used generative AI during a recent purchase, primarily to gather information about vendors and products. Those buyers used an average of seven different information sources during the purchase process.
That gives us a better picture of what may be happening.
AI does not necessarily replace the buying journey. It becomes part of it.
AI may build the shortlist. The human conversation confirms it.
There is another finding in that Gartner research that I think is especially relevant to tax and accounting firms.
Sixty-nine percent of B2B buyers said they prefer to validate AI-generated insights with a sales representative.
For an accounting firm, translate “sales representative” into something more familiar: the consultation call, discovery meeting or conversation with a partner.
The prospect may use AI to understand the problem, identify possible firms and narrow the field. Then they want a human being to tell them, “Yes, we understand your situation, and here's how we would approach it.”
That is why I don't believe AI eliminates the website or the relationship. It changes when they enter the process.
Instead of arriving on your homepage knowing almost nothing about you, a prospect may arrive after AI has already researched your firm, compared it with others and formed an initial impression. Your website and your first conversation increasingly become places where the prospect validates what AI has already told them.
That is a very different buyer journey.
This shift may favor smaller specialized firms
For a 10- or 15-person accounting firm, this may be one of the most important parts of the story.
You are probably not going to outspend a national firm on advertising. You may not publish dozens of articles every month. You probably do not have an internal SEO department.
But AI discovery is not only about scale. It is also about relevance.
A prospective client can ask:
I run an $8 million construction company and expect to sell it within five years. I need a CPA who understands tax planning and business transition.
Or:
I'm a dentist planning to sell my practice and retire in three years. What kind of tax advisor should I be looking for?
Or:
I own several rental properties and may do a 1031 exchange. Which firms near me specialize in real estate taxation?
Those are far narrower questions than “CPA near me.”
That creates an opportunity for a smaller firm with real depth in a niche. A 12-person firm that clearly demonstrates expertise in construction, dental practices, real estate or closely held business transitions may be more relevant to a very specific question than a much larger regional firm with generic positioning.
The advantage comes from specificity.
The problem is that many firms hide that specificity.
A physician-focused firm with twenty years of experience might still describe itself this way:
We provide personalized tax and accounting services to individuals and businesses.
That sentence could belong to thousands of firms.
A clearer version might say:
We help physicians and medical practice owners reduce tax exposure, improve practice profitability and plan for major transitions such as a practice sale or retirement.
The second version gives a human prospect more information. It also gives an AI system much more evidence about who the firm serves and what problems it understands.
That is the real goal of AI visibility: making genuine expertise easier to identify.
Your website is only part of the evidence
This is also where firms can make a mistake.
They assume AI visibility is something that happens entirely on their website.
It isn't.
Remember Gartner's finding that B2B buyers used an average of seven information sources during a recent purchase. Prospects, and the AI systems helping them research, can encounter information about your firm in many places beyond your homepage.
That broader footprint may include reviews, professional directories, state society listings, association profiles, podcast appearances, local media, guest articles and other credible third-party sources.
If your website says you are a leading advisor to construction companies, that is one piece of evidence. If clients describe your construction expertise in detailed reviews, you speak on construction tax issues, industry organizations list you as a resource and other websites cite your work, the picture becomes much stronger.
Your authority lives across the web.
Technical structure matters too. In How Schema Helps You Show Up in AI Search, Not Just Google, we explain how structured data can make the people, services, locations and expertise represented on a firm's website easier for machines to interpret.
But schema alone won't create authority that isn't there. The broader objective is to make your real-world expertise visible in enough credible places that both people and machines can understand it.
So how do you actually measure AI visibility?
This is the part I think most firms are missing.
You do not need to begin with an expensive technology stack. You can build a useful baseline manually.
Start with 15 to 25 questions that your ideal clients might actually ask. Do not write them like SEO keywords. Write them the way a real business owner would talk.
A construction-focused firm might use prompts such as:
I own a $5 million construction company and need more proactive tax planning. Which accounting firms near me specialize in contractors?
I'm thinking about selling my construction business in the next five years. Which CPA firms can help me prepare for the sale?
My construction company has grown quickly and I'm not sure my current entity structure still makes sense. Who should I talk to?
A real estate-focused firm would build a completely different set.
The key is that the prompts should reflect your actual growth strategy. If you want more medical practices, measuring whether AI recommends you to restaurants doesn't tell you anything useful.
Then run the same prompt set across the major AI discovery environments you care about, such as ChatGPT, Gemini, Perplexity and Google's AI-powered search experiences.
You are looking for four things.
1. How often are you mentioned?
If you run 20 relevant prompts and your firm appears in two answers, your starting visibility rate is 10%. If you appear in 12, it is 60%.
I would not obsess over an industry benchmark yet because this category is still young and results vary by geography, platform and query. The useful benchmark is your own baseline and whether it improves over time.
You can also look at share of voice by comparing how often your firm appears versus the same competitors across the same set of prompts.
2. Which competitors appear when you don't?
This may be even more useful than your own score.
If another local firm consistently appears for the clients you want, study what the web says about them. Look at their service pages, niche content, reviews, association visibility, media mentions and third-party references.
That gives you a more useful competitive question than simply asking who ranks above you for one keyword.
3. Which sources does AI rely on?
When citations or source links are available, record them.
You may find that AI repeatedly relies on a professional directory, an association website, a local publication, a detailed article, reviews or a competitor's site.
That tells you where authority is being established.
If another firm's article is repeatedly cited on an issue where your firm has deep expertise, that is not simply a search ranking problem. It is evidence that the internet has a stronger body of material connecting that firm with the subject.
4. Does AI describe your firm accurately?
Visibility is not automatically good visibility.
AI might find your firm but describe it primarily as a tax preparation practice when advisory is your growth strategy. It may misunderstand your specialties, use outdated information or fail to recognize the markets you serve.
Record that as well.
The goal is not simply to be mentioned. It is to be understood correctly.
Use a consistent methodology
There is an important caveat here.
AI answers are not static search results. They can vary from one run to another and can be influenced by location, conversation history, account memory and changes in the underlying models.
So don't run one prompt once and call the result your “AI ranking.”
For a more defensible baseline, use a fresh or logged-out session where practical, keep the location stated in the prompt consistent, and run each important prompt two or three times. If you're calculating a visibility score, count a firm as meaningfully present when the mention recurs rather than relying on one isolated answer.
The objective is not laboratory-level precision. It is consistency.
Use the same process each time and watch the trend.
Put it on a simple scorecard
You don't need a 40-page report.
A basic scorecard could look like this:
Run the same core prompt set again later and compare.
For most small and midsized firms, monthly tracking is probably enough if AI visibility is an active growth initiative. Quarterly may be sufficient if you simply want to monitor how the market is changing.
What I would not do is run a prompt once, see your firm appear and declare victory. Sources change. Competitors publish. Reviews accumulate. Models change.
The trend matters more than any single screenshot.
You can absolutely do this manually with a spreadsheet when you are starting. Once you have dozens of prompts, multiple locations, several service lines and multiple AI platforms, software begins to make more sense because manual testing becomes tedious quickly.
But the principle does not change:
Measure whether the market you want is increasingly seeing your firm inside AI-generated recommendations.
That is much more useful than simply saying, “We need to do GEO.”
The traffic is still small compared with Google. That's not the point.
It is worth keeping the scale in perspective.
Similarweb's September 2026 analysis of AI referral traffic found that AI platforms generated an average of 770.7 million referral visits per month worldwide between June 2025 and May 2026, up 117.4% year over year.
Traditional search still sends far more traffic.
So this is not an argument to abandon SEO or suddenly move your entire marketing budget into AI optimization. The reason to measure AI now is not because it has already replaced Google. It is because the channel is growing rapidly, buyer behavior is changing and firms now have another meaningful way to understand whether they are visible during the research process.
Traditional search visibility still matters. Local rankings still matter. Reviews still matter. Your website still matters.
The definition of visibility has simply expanded.
A one-hour starting point
Take your five most valuable client types and write down three or four questions each of those clients might ask before hiring an accountant.
Run those questions through the major AI platforms and record who appears.
Then go one level deeper. What sources are being cited? What does AI understand about the firms it recommends? What expertise does it associate with them? What evidence exists about them across the web that may not exist about you?
That is where AI visibility becomes useful.
You stop debating whether GEO is real and start measuring whether your firm is part of the conversation.
For the last twenty years, tax and accounting firms have spent enormous amounts of time asking:
Can prospective clients find us?
The next question belongs right beside it:
When AI is asked who to hire, will it recommend us?
See How Ready Your Firm Is for the Next Search Era
AI visibility is one of the areas we look at when evaluating whether a firm is positioned for what's next. But it does not exist in isolation. Technology, AI adoption, client experience, visibility and growth strategy increasingly overlap.
The Modern Practice Score looks across those areas and gives our team the starting point for your personalized Future-Ready Firm Blueprint.
You'll see where your firm is ahead, where gaps are forming and which changes deserve attention first.
Get My Future-Ready Firm Blueprint →
