How do accountants get customers from ChatGPT?

How does the owner journey run through AI?

A business owner does not start by looking for an accountant. They start with a problem and ask the engine to settle it. The journey runs in three moves. First the threshold question: "do I need a CPA or is a bookkeeper enough," or "when should a freelancer switch from TurboTax to an accountant." This decides whether they hire anyone.

Then the cost question: "what does a CPA cost for an S-corp," or what a monthly bookkeeping engagement runs. This sets expectations and filters the field. Only then comes the hiring question: "CPA for small construction businesses in {city}." A firm that shows up only at the last step arrives after the owner has already formed opinions from whatever sources answered the first two. The firms that get the call are usually the ones the engine cited at the threshold.

The reason this matters is that the early questions shape the later ones. An owner who learned from the engine that an S-corp election is worth it carries that frame into the hiring question, and looks for a firm that handles S-corps. An owner who was told a bookkeeper is enough never asks the hiring question for a CPA at all. By the time the buyer is ready to choose, the engine has already narrowed the field on their behalf, using whatever sources it trusted at each step. A firm that was absent for the threshold and cost questions is competing for a decision that was partly made without it.

Which firms actually get named?

Two things decide it: verification and specialization. Verification is unusually concrete for accounting, because CPA licenses are public. State boards of accountancy publish them and CPAverify.org aggregates them, so an engine can confirm a firm is licensed before it names one. A firm whose credentials check out, and whose name and details read the same across its site, the state board, and the directories, is one the engine names with confidence.

Specialization is what matches you to the question. An engine answering "who handles restaurant bookkeeping" looks for a source that says, in readable text, that it works with restaurants. "We work with construction contractors and restaurants" is something the engine can lift and match. "Full-service accounting solutions" matches nothing. The firms that get named state who they serve plainly and back it with verifiable credentials.

The two signals reinforce each other. Verification tells the engine the firm is real and licensed; specialization tells it the firm fits this particular question. A firm strong on one and weak on the other rarely gets named: a verified generalist with no stated focus loses the industry-specific questions, and a firm that claims a focus the engine cannot confirm against a public license loses the trust that earns the citation. The firms that win consistently are the ones that are both confirmable and specific, because that combination is what lets the engine name them without hedging.

What are the steps to get your firm named?

1. Run the check on the threshold questions in your specializations. Ask ChatGPT, Claude, Gemini, and Perplexity the questions your prospects ask first, such as whether they need a CPA and who handles their industry, then record which firms get named and which sites get cited.
2. Publish one plain page per question. Write a page that answers a single question directly: when a freelancer needs an accountant, what an S-corp return costs in range terms, when software stops being enough. One question per page, in text an engine can quote.
3. Name your industries in readable text. Say which businesses you serve in plain sentences, such as construction contractors or restaurants, so an engine can match your firm to an industry-specific question instead of guessing.
4. Keep your details identical everywhere. Make your firm name, credentials, and services read the same across Google, your state-board listing, and your site, so the engine can verify you without hitting conflicting information.
5. Add fresh dates and credential schema. Show a clear last-updated date on each page and mark up your firm and its credentials in structured data, so the engine can read who you are and confirm you are current.
6. Re-check in two to four weeks. Ask the same questions again and compare. Watch which answers added your firm and which open questions you claimed, then publish the next page against whatever is still missing.

The loop is plain enough to run yourself. The work is choosing the right questions, writing each page so an engine can quote it, and keeping the verification clean.

What actually moves AI visibility?

The page features that earn citations are measurable, and they line up with how a firm should write. A study from Princeton tested which changes increased how often AI assistants cited a page, and the pattern favors exactly the plain, sourced writing a licensed firm can produce.

For an accounting firm this is a fit, not a stretch. A page that cites the relevant rule, states a cost range with the figure attached, and explains the trade-off in a quotable sentence is the page the engines reward. The same writing that keyword stuffing penalizes is writing no good firm wanted to publish anyway.

The takeaway is that the work rewards substance, not volume. A firm does not need to publish constantly or chase keywords; it needs to write a small number of pages that actually answer the questions, with the specifics an engine can quote. State the range and what moves it. Attribute the rule you are explaining. Use the words a real owner would use rather than a category label. Those are habits a firm already has in client conversations, which is why the firms that struggle with this usually struggle with publishing at all, not with knowing the answer.

How do you time the work against tax season?

Publish before the question wave, not during it. Tax season concentrates the questions, but the engines need time to crawl a page, index it, and start trusting it. A page published in February is competing from behind while demand peaks. A page that has been live and consistent since the prior autumn is the one already being cited when the wave arrives.

The threshold and cost questions are evergreen, so they reward early publishing the most. Entity-formation, payroll, and bookkeeping questions run year-round, which means the work is never wasted out of season. Treat the calendar as a deadline that runs months ahead of the spike: the pages that win tax season are the ones you published well before it.

There is a quieter benefit to publishing in the off-season. Summer and early fall are when a firm has the time to write carefully, and they are also when the year-round questions are easiest to claim because fewer firms are paying attention. A page about when a freelancer needs an accountant or whether to elect S-corp status faces less competition for the citation in August than the same page would in March. By the time the seasonal owners arrive with their questions, the engine has already decided which sources it trusts, and an off-season page has had months to become one of them.

How do you know it is working?

You know by re-running the same questions and watching the answers change. The measurement is the proof. Track two outcomes per question: whether your firm is named in the answer and whether your site is cited as a source. A name reflects reputation the engine absorbed; a citation means your own page was read and quoted. Both matter, and a real check counts both.

Watch for the citation before the name. When a page first starts working, the engine begins citing it as a source on the question it answers, sometimes before it names your firm in the answer text. That early citation is the leading signal that the page is doing its job, and it tells you the approach is sound even if the named-mention count has not moved yet. As the page accumulates trust and the verification signals line up, the named mentions follow. A firm that tracks both outcomes per question can tell the difference between a page that is being read and a page the engine has not found, which is exactly the information needed to decide whether to wait, revise, or write the next page.

Tenva runs this loop as a service and as a free starting check. We put your firm through the four-engine measurement, show you every answer, and tell you which open questions to claim first. Start with the free AI visibility checker to see where your firm stands today.