While word of mouth, digital marketing and professional networks are still the main ways accounting firms promote themselves, an entirely new path has emerged over the past few years: AI. With more people using AI every day, CPA firms across the country are reporting an uptick in clients who discovered them through conversations with ChatGPT, Gemini, Claude and other models.
AI-sourced client referrals are a relatively recent phenomenon, with most firms only starting to get them this year or last. This makes sense, as many early AI models lacked the ability to browse the internet, and their overall user base was smaller. But now there are hundreds of millions of people using AI every day, some of whom need accountants.
In general, these clients are not asking their AI models to find them a CPA, any CPA, and picking from a list of suggestions. Rather, they tend to query their models on highly specific accounting or finance issues, and then find the firms through the supporting links provided. For example, Patrick Camuso, head of crypto-specialist firm Camuso CPA, reported that most people who contact him through AI have specific needs related to digital assets.
"It will be 'I got a tax notice and haven't done my crypto accounting for five years' or 'I need crypto tax planning for my portfolio' or 'I need to set up a Web3 accounting system.' It can vary with what they type in. Sometimes it's broad-based crypto [topics], sometimes it is very specific with what they're looking for. But it is all basically crypto related," he said.
Sasha Tchulkova, marketing director of advisory services for Top 25 firm Withum, says her firm has had similar experiences. One recent example was a venture capital firm with what she said was a "very unique scenario" that, at a certain point in a deal, required "very specific accounting support with the necessary compliance and accounting needs." The client used ChatGPT to evaluate Withum's ability to address the issue versus other firms.
"They went into ChatGPT and put exactly their scenario in," said Tchulkova. "They already had Withum on the radar, so they [then] asked to compare Withum with three or four other CPAs based on their needs and what would meet them. After reviewing that, they knew Withum was it, so when we had a sales call they had already decided who they were going with."
The hyper-specificity of these clients was noted by other firms as well. For instance, while her practice offers many different services, from tax advisory to bookkeeping, Katherine Bunschoten, head of North Carolina-based Certum Solutions, noted that pretty much all the clients referred to them by AI wanted the same thing: accounting software training.
"It's not so much straight bookkeeping they're really looking for. … A lot of them are trying to find how to do things within their software and then they're running into some of our content," she said.
Similarly, while Withum also offers a wide variety of services, Tanina Khanuja, the firm's digital marketing director, said that every AI-referred client that contacts them needs tax services.
"Right now, the clients we acquired through AI are all tax services… When we do attribution tracking, right now it is tax services all the time," she said.
Sometimes this specificity is less about type of service and more about sector expertise. Suzanne Reed, chief marketing officer for top 50 firm LBMC, agreed that people aren't asking for a general CPA and instead are making targeted inquiries on specific issues, some of which her firm happens to specialize in.
"Health care continues to be the top driver, especially around services like health care valuations, physician compensation, MSO structuring and broader health care consulting. We've also seen interest in cybersecurity audits and related compliance work. These prospects typically know what they need, but they're not sure who to call — so they're turning to AI to point them in the right direction," she said.
Another point mentioned by multiple firms is that leads from AI tend to be of high quality, perhaps because of this specificity. Camuso, for example, noted that clients referred to him by AI tend to already be familiar with blockchain and cryptocurrency concepts, which significantly eases onboarding, as he does not need to spend time educating them and developing trust. They usually come in eager to work with him.
"[They're] very motivated to work with me already. It's very contextual. They're like 'Hey, I'm looking for X, Y or Z. Can you confirm if you can do this?' It is very calibrated. None of these were wild random leads, which is very interesting," he said, adding that they also tend to do their own research.
Kathleen O'Toole, chief marketing officer for Top 25 firm PKF O'Connor Davies, reported similarly, noting that the ratio of qualified leads over total leads for such referrals is actually higher than the firm-wide ratio. Further, she said, the conversion rate from AI traffic is three times that of the site average (with the top referrer being, by far, ChatGPT), which is likely fueled by the specific content on the firm website.
"We can see in our reporting on web leads that they came to our site and in most cases landed on a specific service or industry page that had content that was apparently relevant to their conversation," she said.
Another point of commonality mentioned over and over is that these AI-referred clients tend to be younger and more tech savvy, which aligns with the overall demographics of AI users.
"[They're] definitely more tech focused, for sure. These are people who are embracing new technology, much like ourselves, and are learning how much artificial intelligence can help in the right context," said Bunschoten.
Despite these strong commonalities, there are wild inconsistencies from firm to firm. For one, the number of AI-referred leads varies greatly. Camuso said AI has already become a serious referral channel for his firm, as at this point such clients make up 15-20% of his overall leads. Similarly, Bunschoten said AI referrals have become "pretty common" at her firm, now making up between 10-20% of her overall leads.
Conversely, Reed from LBMC said that while the firm has noticed a clear increase in AI-attributed leads over the last several months ( AI traffic has increased 200% in the past year), they remain a small slice of LBMC's overall leads, an estimated 5-7% (though Reed noted it was 0 last year). Similarly, Khanuja, from Withum, said 1% of the total traffic comes from AI and, of that, 0.5% of conversions are from AI.
There's also little consistency where these leads come from. Camuso, Bunschoten and Reed all said their referrals are from all over the world. However Khanuja and Tchulkova, from Withum, said the referrals tend to come from Withum's primary markets. And O'Toole from PKFOD said about 40% of its AI referrals came from either New York or New Jersey, "which makes sense because we have more offices there and we are headquartered in NY."
Regardless of where AI referrals lead, there was broad agreement that firms will likely be seeing more of them as time goes on, whether it's a steady climb or a dramatic spike. This means that understanding AI and how it chooses what information to convey will become increasingly important for accounting firms looking to get noticed.
"I do believe that more is going to get funneled into AI … It's probably going to get more competitive, so it's hard to say exactly [how it will work for everyone] but I think [the number of] people using AI is going to go up and the people that go and find service providers and CPAs are going to be doing it with these AI systems," said Camuso.
But wait! How did the AI models find these firms to begin with? And what can firms do now to encourage AI-guided recommendations? Read on in part 2 of this story!
