Key takeaways
- We ran our live AI visibility audit tool against 23 established accounting and CPA firm sites and could scan 15. Average visibility was 32.0% across 75 live, grounded buyer queries. The median firm scored 20%.
- Thirteen of 15 firms were named in two of five queries or fewer. Two were not named once, and the best score was 80%, reached by BDO and RSM US.
- Accounting answers look like ranked lists. The comparison query ("top accounting firms") did best at 53.3%. The query where an owner describes a problem named a firm 20.0% of the time, and the local query only 13.3%.
- The biggest brands crowded the answers. Deloitte was named as an alternative in 17 of 75 answers, Grant Thornton in 12 and PwC in 11, even on mid-market questions.
- A firm's own website was 4 of 584 citations (0.7%). AI leaned on Wikipedia, Yelp, Clutch, vendor-written ranking lists, and a few firms' city-by-city location pages.
- Eight of 23 sites could not be scanned, which is notable. We cannot tell why or whether AI crawlers are affected. The sample is mostly large firms and likely flatters smaller ones.
Why we audited accounting firms
Accounting is a trust business that grows by referral. A lawyer, a banker or a friend with a similar company says "call these people", and for decades that was enough. It still works. What has changed is what the referred owner does next, and what the owner with no referral does at all. Both now ask an AI assistant.
The moments that bring an owner to a firm are specific. A lender asks for audited numbers. A buyer appears. The books are a mess, or an accountant has gone quiet. The owner describes that situation to an assistant and reads the shortlist. So we asked: when a buyer describes an accounting need without naming a firm, which firms get named, and is being large and well known enough?
32.0%
average AI visibility across 15 established accounting firms, measured by live grounded search on 75 buyer queries
Deligatr Accounting Firm AI Visibility Audit, October 3, 2026
How the audit works
The tool reads each firm's homepage, classifies it in the words a buyer would use, writes five buyer-intent queries (one local-style, two category, one problem-first, one comparison) and runs each as a live, grounded AI search with the firm's own name left out. It is one run, five queries per firm, on one day. AI answers change between runs, so read every score as a snapshot. The full method is in our audit documentation, and all twelve industry audits are on our AI visibility research page.
A firm is either in the answer or it is not.
The distribution: only two firms reached 80%
Firms by AI visibility score
15 established accounting firms, each tested on 5 live grounded buyer queries. Number of firms at each score.
Deligatr Accounting Firm AI Visibility Audit, October 3, 2026
BDO and RSM US led at 80%. Five firms sat at 40%: Cherry Bekaert, CohnReznick, Forvis Mazars US, Grant Thornton and Plante Moran. Six scored 20%, including CliftonLarsonAllen, Citrin Cooperman and Aprio. EisnerAmper and Withum were not named in any of their five queries.
A score reflects five specific queries, and queries differ by firm. EisnerAmper scored 0% on its own five, yet it was named as an alternative in 7 of 75 answers to other firms' queries. Treat each score as a snapshot of those five questions, not as a ranking of the firms.
What kind of question loses the most
How often a firm was named, by query type
Share of queries in which the firm was named. 15 firms, 75 queries.
Deligatr Accounting Firm AI Visibility Audit, October 3, 2026
Accounting answers behave like ranked lists. When the question was "top accounting and advisory firms" or "biggest firms in the US", eight of 15 firms were named. Those are the questions where size and brand recognition count, and the firms in this sample are big.
The queries that sound like a real buyer did worse. Problem-first queries such as "our company needs help with ERC tax credits and IRS compliance" or "we need help with our annual audit and tax strategy, what kind of firm should we hire" named a firm 3 times in 15: BDO, RSM US and CohnReznick. Local queries named a firm twice, Cherry Bekaert and Rehmann. One caution: several local queries were a generic "near me" with no city, so treat that number as the roughest of the four.
The pattern differs from our law firm audit, where the generic category query did worst, and from our insurance brokerage audit, where the problem-first query did worst but the local query was stronger.
Who got named instead, and where the answers came from
When a firm was not named, AI named someone else, and it was usually the same few brands. Deloitte appeared as a named alternative in 17 of 75 answers. Grant Thornton appeared in 12, PwC in 11, BDO USA in 10, RSM US LLP in 9, and KPMG and Cherry Bekaert in 8 each. EisnerAmper and Baker Tilly each appeared in 7. For a set of mostly mid-market questions, the Big Four were still prominent.
The sources behind the answers were fragmented. Across 584 citations on 273 domains, Wikipedia was the largest at 54, which is 9.2% of the total. Much of that was attention on a few firms. The Moss Adams page was cited 15 times, LaPorte 10 and Aprio 8.
Three other groups stand out.
Review and directory sites. Yelp was cited 19 times, 16 of them on local-style queries. Clutch was cited 15 times, mostly on category and comparison queries.
Ranking lists written by vendors. ACE Cloud Hosting (14), A-One Outsourcing (11), Zamp (8) and Cloudvara (7) together made up 40 citations. These appear to be hosting, outsourcing and software vendors publishing "top accounting firms" lists, including city-by-city versions. Together they were cited more often than any firm's own site.
Other firms' location pages. The most striking pattern was a handful of firms' city pages cited repeatedly on local queries. Kemper CPA Group, which was not in our sample, was cited 15 times, all on local-style queries, all to city location pages. Cherry Bekaert's cbh.com was cited 11 times, 8 of them on local queries, with city pages for Dallas, Houston, Tampa, Raleigh, Greenville, Washington and Austin. AI did name firms for local questions, but it drew on pages that said something specific about a place.
4 of 584
citations pointed to the firm's own website, 0.7%, measured by matching the cited domain exactly to the firm's own domain
Deligatr Accounting Firm AI Visibility Audit, October 3, 2026
Those four were Weaver, CohnReznick, BDO and RSM US, each pointing to a service or practice page. One firm in the sample, Cherry Bekaert, publishes at a second domain (cbh.com), which our exact-match rule does not count as its own. Counting that one citation would make it 5 of 584, 0.9%. Either way, the point stands. AI built its answers from lists, directories and reference sites, and from a few firms' location pages. A strong website does not make up for a thin presence in those places.
The eight we could not scan
We attempted 23 sites and could scan 15. The eight that could not be audited were Armanino, Baker Tilly, CBIZ, Crowe, Friedman, Marcum, Moss Adams and Sax. In each case the tool could not read enough of the site to run. That is about a third of the sites we tried, which is notable for a profession that handles sensitive data and may well run strict bot protection.
What we cannot conclude is why, or whether the same protection stops the crawlers AI assistants use. A failed scan is not the same as being invisible. Baker Tilly was named as an alternative in 7 of 75 answers and bakertilly.com was cited 7 times. Moss Adams's Wikipedia page was cited 15 times, and Armanino's site 5 times. AI had plenty to say about firms our tool could not read. The reverse risk is also real. If a firm's protection does block AI crawlers, its own content is not being read at all, and you would not know unless you tested.
What the labels tell us
Our tool read 9 of the 15 firms as a general "accounting and tax advisory firm" or "accounting and business consulting firm". The most specific label in the sample was "audit, tax and advisory firm for mid-market and enterprise businesses". None was read as an industry specialist, even though several of the firms are known for sector depth.
We did not run a controlled test, so we are not claiming that specialization causes visibility. These are large, multi-industry firms and the classifier only read a homepage. The point for a smaller firm is practical. AI answers a question about an industry, a service and a place. A firm that states, specifically and consistently, what it does for whom is easier to place than one that describes itself as full service.
What to do with this, if you run a firm
- Say what you are best at in one plain sentence. "CPA firm for manufacturers and distributors in Ohio, with R&D credit and outsourced CFO services" is easy to place. "Full-service tax, audit and advisory" is not. List the industries, services and software platforms you support, on pages of their own.
- Publish the decisions, not only the deadlines. Whether you need a CPA or a bookkeeper, what an audit costs and what drives it, when an outsourced CFO makes sense, what a quality of earnings review involves. These are the questions problem-first queries ask, and few firms answer them plainly. Keep any seasonal tax pages current each year.
- Build location pages that say something real. The firms cited on local queries had a page per city. Thin copy with the city name swapped is not what you want. Include the team, the industries served there and real local detail.
- Treat directories and lists as part of your marketing. Check your name, offices, services and partner profiles on Yelp, Clutch, your state society's directory and any rankings relevant to you. Inconsistent entries make you hard to place.
- Be careful with claims. AICPA rules do not allow false, misleading or deceptive advertising, and referral fees and commissions must be disclosed. Independence rules affect what you can say about attest clients. Your state board's rules apply on top.
- Check whether your site can be read, then re-test. Test what your security layer shows a plain automated visitor. AI answers change, so a single score tells you little.
If you want this done as a program, that is what our AEO, GEO and SEO service for accounting and CPA firms is built for. We also have a case study on a referral-led accounting firm's first outbound campaign, which is one client's experience and not a typical result.
Every firm we audited
Firm | Visibility | Appeared on | Top rival named instead |
|---|---|---|---|
| BDO | 80% | Category x2, Problem, Comparison | Deloitte |
| RSM US | 80% | Category x2, Problem, Comparison | Grant Thornton |
| Cherry Bekaert | 40% | Local, Comparison | RSM US LLP |
| CohnReznick | 40% | Category, Problem | RSM |
| Forvis Mazars US | 40% | Category, Comparison | CBIZ |
| Grant Thornton | 40% | Category, Comparison | PwC |
| Plante Moran | 40% | Category, Comparison | EisnerAmper |
| Anchin | 20% | Comparison | George Dimov CPA |
| Aprio | 20% | Category | Cherry Bekaert |
| Citrin Cooperman | 20% | Category | EisnerAmper |
| CliftonLarsonAllen | 20% | Comparison | Cherry Bekaert |
| Rehmann | 20% | Local | BDO USA |
| Weaver | 20% | Category | Cherry Bekaert |
| EisnerAmper | 0% | None | Deloitte |
| Withum | 0% | None | Baker Tilly |
Not scanned (the tool could not read enough of the site): Armanino, Baker Tilly, CBIZ, Crowe, Friedman, Marcum, Moss Adams, Sax.
Check your own number
Every firm above was run through the same tool below. It takes about ninety seconds, and your firm's name is never in the query.
Do AI assistants recommend you?
Five buyer-style queries for your category, run through live AI search, show who gets named instead of you. Takes about ninety seconds.
Free. The full report arrives by email so the tab can close.
Methodology and honest limitations
What we did. On October 3, 2026 we ran the live audit tool that powers our free checker against 23 established accounting and CPA firm sites: one run, five queries per firm, visibility measurement only. We could scan 15 of the 23. The step-by-step method is in our audit documentation. A firm that disputes its result can ask us to re-run it.
The sample. We chose real, identifiable firms, mostly large national and top-tier regional firms, picked by us and not at random. It is not a census and not representative of small or local practices. It likely flatters smaller firms, because large firms have far more written about them, so smaller firms may fare worse than these did. A firm's score also depends on how our classifier read its homepage.
What it does not cover. Only the homepage was read, so a firm with a thin homepage and strong inner pages may be undercounted. Five queries per firm is a small sample of real buyer language, and several were broad rather than tied to one industry or city. Results come from one run on one day, and AI answers change over time. Eight of the 23 sites could not be scanned, so the results exclude them, and we do not know whether they would have scored differently. We counted a citation as the firm's own site only when the cited host exactly matched the firm's domain or a subdomain of it, and checked those matches by hand. We did not test whether any firm pays for directory or list placement.
On professional rules. Nothing here is professional or legal advice. Rules on CPA advertising, referral fees, independence and firm names vary by state, and your own state board of accountancy and the AICPA Code of Professional Conduct apply.
FAQ
How visible are established accounting firms in AI answers?
Not very. Across the 15 firms we could scan, average visibility was 32.0% over 75 buyer queries. Thirteen firms were named in two of five queries or fewer, two were not named at all, and only BDO and RSM US reached 80%.
Which firms did AI name instead?
The largest brands, even for mid-market questions. Deloitte was named as an alternative in 17 of 75 answers, Grant Thornton in 12, PwC in 11, BDO USA in 10, RSM US LLP in 9 and KPMG and Cherry Bekaert in 8 each.
Where does AI get its answers about accounting firms?
From many places, and rarely the firm's own website. Across 584 citations on 273 domains, Wikipedia was the biggest source at 54, followed by Yelp, Clutch and a mix of ranking lists and individual firms' location pages. A firm's own website was 4 of 584.
Does this audit tell us how small or regional firms fare?
No. The sample was mostly large national and top-tier regional firms that we chose, not a random sample. It likely flatters smaller firms, which have less written about them. Treat the numbers as a picture of established, well-known firms.
Why could eight sites not be scanned?
We attempted 23 sites. For eight of them, our audit tool could not read enough of the site to run. We cannot tell from outside whether the cause was bot protection or something else, or whether it affects the crawlers AI assistants use. We draw no conclusion from it beyond noting that about a third of the sites we tried did not scan.
Can a firm pay for a better result?
Not in a way we would recommend. Some directories and ranking lists sell placement, and we cannot see how AI weighs that. The reliable route is accurate profiles, clear pages on the industries and services you specialize in, and genuine third-party mentions, within your state board and AICPA rules.
Related reading
- AEO, GEO and SEO for accounting and CPA firms
- We audited 22 established law firms for AI visibility
- We audited 16 established insurance brokerages for AI visibility
- How to get cited in ChatGPT, Perplexity and AI Overviews
- What is answer engine optimization (AEO)? The complete guide
- AI visibility research: every audit in one place
