Accounting artificial intelligence leaders have deemed Chinese large language model DeepSeek as roughly on par with other general models when it comes to knowledge, questions and tasks relevant to the profession.
DeepSeek’s latest R1 model was released to the world last week to much fanfare by producing performance comparable to massive models like Claude or ChatGPT but at a fraction of the cost — something that blindsided other players in the AI field who hadn’t believed it was possible. In the short time since it exploded on the world stage, industry observers have had to rethink their priors, as the event has shown that one does not have to be a gigantic corporation like Microsoft (whose $3 trillion market cap is bigger than the entire GDP of Italy) to produce quality AI models.
Generative AI models are not usually known for their math skills, let alone accounting, as the probabilistic nature of their outputs makes precision difficult. While theoretically someone could apply public models like ChatGPT to an accounting problem, the possibility the AI could make up information from whole cloth makes it a risky proposition. This does not, however, mean they’re completely incapable of performing accounting tasks, just not as good as a more specialized solution.
Alexandr Vasilyev – stock.adobe.
Jeff Seibert, CEO of accounting automation solutions provider Digits, tested DeepSeek against OpenAI’s ChatGPT, Anthropic’s Claude and Meta’s Llama and found its ability to perform accounting tasks and answer relevant questions to be roughly on par with the others. He asked the AIs to classify 1000 transactions, given the description, dollar amount and a chart of accounts. ChatGPT was correct 61-65% of the time depending on the specific version; DeepSeek was correct 59.90% of the time; Llama was correct 48% of the time; and Claude was correct 43.4% of the time. The correct category was in the top 3 suggestions 75-78% of the time for ChatGPT (depending on version), 69.67% for DeepSeek, 61% of the time for Llama and 56% of the time for Claude.
ChatGPT created a fake category 0.10-4% of the time (depending on version), Llama did so 0.2% of the time, Claude 2.8% of the time, and DeepSeek 0.39% of the time. As far as speed, it answered queries faster than GPT-o1 mini, GPT-o1, and Llama. It was slower than GPT-4o, GPT 4 Turbo and Claude.
Seibert, in a LinkedIn post outlining his experiment, noted that even if it doesn’t outdo the other models in all areas, the fact that it was made at a fraction of their cost is quite impressive.
“If their claims around training cost are accurate, this represents a massive breakthrough in model efficiency and sets the new bar for open source AI performance,” he wrote.
Daniel Shorstein, president of technology solutions advisory firm James Moore Digital, also put DeepSeek through its paces by asking multiple choice questions covering a range of accounting topics, adding he tried to keep them just difficult enough that it would trip up a less than highly intelligent LLM. He used the same test as the one he made to evaluate Llama, Claude, ChatGPT and other models against each other. He found its results to be inconsistent.
He illustrated with how it reacted to a question on segregation of duties:
Question: “There are three employees in the accounting department: payroll clerk, accounts payable clerk, and accounts receivable clerk. Which one of these employees should not make the daily deposit? A. payroll clerk B. account payable clerk C. accounts receivable clerk D. none (any can make the deposit)”
DeepSeek, like other models lately, is equipped to not only provide an answer but reveal some of its internal reasoning in how it got to the answer. Shorstein noted that, internally, it actually got the correct answer after a long chain of reasoning where it first recalled general principles of segregation of duties, thought of an ideal setup, thought of possible exceptions and special circumstances, went back to the main point of segregation of duties and ultimately determined the answer was C, the accounts receivable clerk, which Shorstein said was the correct answer.
“But its final answer: ‘The correct answer is A. payroll clerk,'” he said in a message.
Meanwhile, Hitendra Patil, CEO of accounting tech consultancy Accountaneur, said DeepSeek gives more detailed answers compared to ChatGPT, and also appears to have been built to not only answer the direct question asked, but to actually pre-empt the likely follow-up question that the user may ask after the first question and provides answers to such pre-empted follow-up questions. He added that it also goes over its reasoning when discussing math questions versus other models which simply give an answer.
For example, he asked the model what is 343 multiplied by 741. It broke down the problem using the distributive property to simplify the multiplication, then added the results together to get 254,163.
At the same time, he noted that DeepSeek does not browse the Internet unless specifically asked, and so some of the answers it gave him about tax law were somewhat dated, versus ChatGPT which more or less gave the latest information. Overall, he said that it seems slightly behind ChatGPT for now, despite rumors that it had been trained similarly.
“There is no verifiable proof … but DeepSeek has been/is being trained on similar underlying data that ChatGPT was/is being trained on, albeit it seems to be lagging behind ChatGPT,” he said in an email.
Accounting firms are reporting bigger profits and more clients, according to a new report.
The report, released Monday by Xero, found that nearly three-quarters(73%) of firms reportedincreased profits over the past year and 56% added new clients thanks to operational efficiency and expanded service offerings.
Some 85% of firms now offer client advisory services, a big spike from 41% in 2023, indicating a strategic shift toward delivering forward-looking financial guidance that clients increasingly expect.
AI adoptionis also reshaping the profession, with80% of firms confident it will positively affect their practice. Currently, the most common use cases for AI include: delivering faster and more responsive client services (33%), enhancing accuracy by reducing bookkeeping and accounting errors (33%), and streamlining workflows through the automation of routine tasks (32%).
“The widespread adoption of AI has been a turning point for the accounting profession, giving accountants an opportunity to scale their impact and take on a more strategic advisory role,” said Ben Richmond, managing director, North America, at Xero, in a statement. “The real value lies not just in working more efficiently, but working smarter, freeing up time to elevate the human element of the profession and in turn, strengthen client relationships.”
Some of the main challenges faced by firms include economic uncertainty (38%), mastering AI (36%) and rising client expectations for strategic advice (35%).
While 85% of firms have embraced cloud platforms, a sizable number still lag behind, missing out on benefits such as easier data access from anywhere (40%) and enhanced security (36%).
Private equity firms have bought five of the top 26 accounting firms in the past three years as they mount a concerted strategy to reshape the industry.
The trend should not come as a surprise. It’s one we’ve seen play out in several industries from health care to insurance, where a combination of low-risk, recurring revenue, scalability and an aging population of owners create a target-rich environment. For small to midsized accounting firms, the trend is exacerbated by a technological revolution that’s truly transforming the way accounting work is done, and a growing talent crisis that is threatening tried-and-true business models.
How will this type of consolidation affect the accounting business, and what do firms and their clients need to be on the lookout for as the marketplace evolves?
Assessing the opportunity… and the risk
First and foremost, accounting firm owners need to be aware of just how desirable they are right now. While there has been some buzz in the industry about the growing presence of private equity firms, most of the activity to date has focused on larger, privately held firms. In fact, when we recently asked tax professionals about their exposure to private equity funding in our 2025 State of Tax Professionals Report, we found that just 5% of firms have actually inked a deal and only 11% said they are planning to look, or are currently looking, for a deal with a private equity firm. Another 8% said they are open to discussion. On the one hand, that’s almost a quarter of firms feeling open to private equity investments in some way. But the lion’s share of respondents — 87% — said they were not interested.
Recent private equity deal volume suggests that the holdouts might change their minds when they have a real offer on the table. According to S&P Global, private equity and venture capital-backed deal value in the accounting, auditing and taxation services sector reached more than $6.3 billion in 2024, the highest level since 2015, and the trend shows no signs of slowing. Firm owners would be wise to start watching this trend to see how it might affect their businesses — whether they are interested in selling or not.
Focus on tech and efficiencies of scale
The reason this trend is so important to everyone in the industry right now is that the private equity firms entering this space are not trying to become accountants. They are looking for profitable exits. And they will do that by seizing on a critical inflection point in the industry that’s making it possible to scale accounting firms more rapidly than ever before by leveraging technology to deliver a much wider range of services at a much lower cost. So, whether your firm is interested in partnering with private equity or dead set on going it alone, the hyperscaling that’s happening throughout the industry will affect you one way or another.
Private equity thrives in fragmented businesses where the ability to roll up companies with complementary skill sets and specialized services creates an outsized growth opportunity. Andrew Dodson, managing partner at Parthenon Capital, recently commented after his firm took a stake in the tax and advisory firm Cherry Bekaert, “We think that for firms to thrive, they need to make investments in people and technology, and, obviously, regulatory adherence, to really differentiate themselves in the market. And that’s going to require scale and capital to do it. That’s what gets us excited.”
Over time, this could reshape the industry’s market dynamics by creating the accounting firm equivalent of the Traveling Wilburys — supergroups capable of delivering a wide range of specialized services that smaller, more narrowly focused firms could never previously deliver. It could also put downward pressure on pricing as these larger, platform-style firms start finding economies of scale to deliver services more cost-effectively.
The technology factor
The great equalizer in all of this is technology. Consistently, when I speak to tax professionals actively working in the market today, their top priorities are increased efficiency, growth and talent. Firms recognize they need to streamline workflows and processes through more effective use of technology, and they are investing heavily in AI, automation and data analytics capabilities to do that. Private equity firms, of course, are also investing in tech as they assemble their tax and accounting dream teams, in many cases raising the bar for the industry.
The question is: Can independent firms leverage technology fast enough to keep up with their deep-pocketed competition?
Many firms believe they can, with some even going so far as to publicly declare their independence. Regardless of the path small to midsized firms take to get there, technology-enabled growth is going to play a key role in the future of the industry. Market dynamics that have been unfolding for the last decade have been accelerated with the introduction of serious investors, and everyone in the industry — large and small — is going to need to up their games to stay competitive.
The House-passed version of President Donald Trump’s massive tax and spending bill would deliver a financial blow to the poorest Americans but be a boon for higher-income households, according to a new analysis from the Congressional Budget Office.
The bottom 10% of households would lose an average of about $1,600 in resources per year, amounting to a 3.9% cut in their income, according to the analysis released Thursday. Those decreases are largely attributable to cuts in the Medicaid health insurance program and food aid through the Supplemental Nutrition Assistance Program.
Households in the highest 10% of incomes would see an average $12,000 boost in resources, amounting to a 2.3% increase in their incomes. Those increases are mainly attributable to reductions in taxes owed, according to the report from the nonpartisan CBO.
Households in the middle of the income distribution would see an increase in resources of $500 to $1,000, or between 0.5% and 0.8% of their income.
The projections are based on the version of the tax legislation that House Republicans passed last month, which includes much of Trump’s economic agenda. The bill would extend tax cuts passed under Trump in 2017 otherwise due to expire at the end of the year and create several new tax breaks. It also imposes new changes to the Medicaid and SNAP programs in an effort to cut spending.
Overall, the legislation would add $2.4 trillion to US deficits over the next 10 years, not accounting for dynamic effects, the CBO previously forecast.
The Senate is considering changes to the legislation including efforts by some Republican senators to scale back cuts to Medicaid.
The projected loss of safety-net resources for low-income families come against the backdrop of higher tariffs, which economists have warned would also disproportionately impact lower-income families. While recent inflation data has shown limited impact from the import duties so far, low-income families tend to spend a larger portion of their income on necessities, such as food, so price increases hit them harder.
The House-passed bill requires that able-bodied individuals without dependents document at least 80 hours of “community engagement” a month, including working a job or participating in an educational program to qualify for Medicaid. It also includes increased costs for health care for enrollees, among other provisions.
More older adults also would have to prove they are working to continue to receive SNAP benefits, also known as food stamps. The legislation helps pay for tax cuts by raising the age for which able bodied adults must work to receive benefits to 64, up from 54. Under the current law, some parents with dependent children under age 18 are exempt from work requirements, but the bill lowers the age for the exemption for dependent children to 7 years old.
The legislation also shifts a portion of the cost for federal food aid onto state governments.
CBO previously estimated that the expanded work requirements on SNAP would reduce participation in the program by roughly 3.2 million people, and more could lose or face a reduction in benefits due to other changes to the program. A separate analysis from the organization found that 7.8 million people would lose health insurance because of the changes to Medicaid.