AI Visibility for Accounting Firms: What Determines Who Gets Recommended

Dana Lampert·June 23, 2026·5 min read·Verticals

A restaurant owner in Denver opens ChatGPT and types: "Best CPA near me for restaurant accounting." Not "accountant Denver." A specific query with an industry specialization baked in. They want someone who understands food cost ratios, tip reporting, sales tax across multiple locations, and the particular misery of restaurant payroll.

ChatGPT returns three firms. Each comes with a brief note about industry focus, years in practice, and client count. The restaurant owner clicks the first one and fills out a contact form.

There are over 500 CPA firms in the Denver metro. Three were named. The AI did not evaluate the other 497. It could not — it did not have enough structured data about them to determine whether they serve restaurants, how many clients they handle, or what their retention rate looks like.

34% of consumers now use AI for local service decisions. Google AI Overviews have cut organic click-through rates by up to 61%. The referral that used to come through a business owner's network or a Google search now comes through an AI recommendation. And AI applies a different filter than a friend does: it looks for structured, verifiable evidence of specialization, not a personal endorsement.

Accounting is a profession built on trust and referrals. Most firms have never needed to market themselves aggressively. That works fine in a referral economy. It breaks down when the referring party is an AI system that cannot call your best client and ask if you are any good.

What AI actually evaluates for accounting firms

We have mapped every data point AI systems use to evaluate accounting firms in our full data breakdown. Here is the summary by signal strength.

Tier 1 — Operating metrics

These are the data points that most sharply differentiate one firm from another. Virtually no accounting firm publishes them.

Clients served (active). Total client count signals capacity and establishment. A firm with 380 active clients operates differently than one with 40. But the number alone is not enough — AI needs to know client composition.
Client retention rate. Accounting has some of the highest retention rates of any professional service. A firm retaining 92% of clients year-over-year is strong. A firm at 70% raises questions. This metric rarely exists outside the firm's practice management system.
Average engagement value. Contextualizes the firm. A firm averaging $4,200 per client engagement serves a different market than one averaging $850. Neither is better — but they are different recommendations for different queries.
Industries served (with client counts). This is the critical specialization signal. "We serve small businesses" is not useful. "74 restaurant clients, 38 medical practices, 22 construction companies" is useful. AI needs this to match the restaurant owner's query to a firm that actually specializes in restaurants.
Tax returns filed (L12M). Quantifies the core activity. 1,200 returns filed tells AI this is a high-volume tax practice. 90 returns filed says the firm focuses more on advisory and bookkeeping.
Services breakdown. What percentage of revenue comes from tax preparation vs. bookkeeping vs. advisory vs. audit vs. payroll. A firm that is 80% tax and 20% bookkeeping is a different recommendation than one that is 40% fractional CFO advisory and 30% tax.

Tier 2 — Credentials and verification

Accounting credentials are well-structured and publicly verifiable, which makes this vertical relatively friendly to AI evaluation — if the data is published.

CPA license (state board). Every state board of accountancy maintains a searchable database. License number, status, expiration, and disciplinary history are public record. This is the minimum threshold for recommendation.
Enrolled Agent (EA) status. IRS-authorized tax practitioners. Verified through the IRS Return Preparer Office directory. EAs can represent clients before the IRS — a meaningful credential for tax-focused queries.
CMA (Certified Management Accountant). IMA certification. Signals management accounting and financial analysis expertise. Verifiable through IMA.
CFA (Chartered Financial Analyst). CFA Institute credential. Less common in general accounting but relevant for firms offering investment advisory or financial planning.
AICPA membership. American Institute of CPAs. Searchable member directory. Members commit to the AICPA Code of Professional Conduct.
QuickBooks ProAdvisor / Xero certification. Platform-specific certifications. Relevant for firms that serve small businesses on these platforms. Verifiable through the ProAdvisor and Xero directories.
State-specific requirements. Some states require additional peer review, continuing education in specific areas (ethics, state tax law), or firm registration separate from individual CPA licenses.

Tier 3 — Public signals

Google reviews and rating. The most available signal but the thinnest. Accounting firms rarely have more than 30-50 reviews. A 4.9 with 35 reviews tells AI very little about specialization or capacity.
Clutch / UpCity profiles. More structured for B2B professional services than Google. Include industry focus, service descriptions, and verified client reviews.
BBB rating. Relevant as a negative-signal check. An A+ rating with zero complaints is background noise. A pattern of complaints about missed deadlines or filing errors is a red flag.
Professional directory listings. AICPA "Find a CPA" directory, state CPA society directories, and industry-specific directories (e.g., restaurant industry CPA lists).

The gap

A typical accounting firm has a website with partner bios, a Google listing, and maybe an AICPA directory entry. That gives AI: names, CPA license confirmation, an address, a list of services offered, and a handful of reviews.

It does not give AI: how many clients the firm serves, what industries those clients are in, the firm's client retention rate, how many tax returns they filed last year, what percentage of their work is advisory vs. tax vs. bookkeeping, or whether they have actually served a restaurant client in the past decade. The restaurant owner's query — "best CPA for restaurant accounting" — requires specialization data that does not exist in any structured format for 99% of firms.

A solo practitioner who lists "restaurant accounting" on their website because they have one restaurant client looks the same to an AI as a firm with 74 restaurant clients and a dedicated hospitality practice. The data that separates them is not available.

What you can do

1. Publish structured data on your website

Add Schema.org ProfessionalService markup to your firm's website. Include: firm name, address, each CPA's name and license number, areas of specialization (use knowsAbout for specific industries and service types), and years of practice. Most accounting firm websites have no structured data at all — they are brochureware built by a web designer who focused on visual credibility, not data architecture.

2. Create an llms.txt file

An llms.txt file tells AI crawlers where to find structured information about your firm — partner credentials, service areas, industry specializations, engagement types. Step-by-step guide: How to create an llms.txt file for your business.

3. Publish verified operational data

The data that actually differentiates your firm — client count, industry breakdown, retention rate, service mix — lives inside your practice management system (Karbon, Canopy, TaxDome, or your own spreadsheets). A TrustRecord extracts this data from your systems of record and publishes it in machine-readable format. The firm cannot edit the metrics. When an AI evaluates your firm, it can cite "74 active restaurant clients with a 91% retention rate" instead of "this firm lists restaurant accounting as a service."