Firms, accountants divided on AI but perhaps not for long

Recent data indicates that the profession is starting to diverge in how it relates to and uses AI, with heavy users on one end and light users on the other, but it may not stay this way for long. 

The AI-Native Firm Advantage Report, a collaboration between the firms KPMG and Baker Tilly plus AI-focused audit solutions platform Fieldguide, found that the split was almost half. The survey, with its 400-person sample, composed of financial auditing and advisory leaders, found 51% of firms were "active deployers" that made AI central to delivery and fully rebuilt their workflows around it to the point where the technology is now integrated across most engagements; conversely, 49% were "casual users" who had used AI to some degree but only in limited or experimental ways and certainly not in a systemic formalized manner across the entire firm. 

The study found that the active deployers significantly out-performed casual users in terms of self-reported profitability increases, capacity expansion, winning clients and being actively sought out by clients, growing advisory revenue, staff retention and the development of client advisory and relationship skills. They were also more likely to have formally redesigned their organizational structure and roles, to have reshaped their career paths, and to have changed their hiring criteria. 

freshidea - stock.adobe.com

While the profession seems divided now, the study's authors feel it will likely shift in the future in favor of the AI-native firms. In an email, a company spokesperson said some firms will deliberately stay in the Casual User camp, using AI at the edges for specific efficiencies because that's what their client base and current model actually need. But apart from these firms, the researchers anticipate the firms adopting AI will build a compounding advantage over others, as Active Deployers are not just adopting tools but building operating models that give them more value over time. That advantage is likely to widen as those firms learn and adapt faster, and embed AI more deeply into how work gets done.

Another study found a similar divergence between accountants. The State of AI in Accounting Firms report from the AI Lab for Accountants, a professional community, found that individual accountants are increasingly split between dabblers and power users with fewer in the middle. Overall, of the sample of 437 people, 45% said they have not gone past dabbling with their main AI assistant; in contrast, 32% of the respondents use AI daily to do things like build custom workflows, projects and Model Context Protocols. 

The report said the distance between the two groups widens every month, though Rebecca Driscoll, one of the founders of the AI Lab for Accountants, noted that since this sample was among those who were already part of the group, it should be read more as a split between a particularly engaged segment of the population. As a result, she was unsure whether this represented a larger split, and if the gap would widen or narrow in the future. However, she noted that proportion of dabblers to power users had a strong relationship with firm resources. 

"It's hard to know for certain. One thing we have noticed is more adoption among larger firms: In our data, the share of power users goes from 10% at solo firms to 35% at firms of 51 to 200 people, and we think that mostly comes down to resources. Smaller firms have the same curiosity, just less time and support to go deep," she said in an email. 

Other studies have also pointed to a widening divide in the profession as well as concerns that the gulf will grow. For example, Intuit's June survey of accounting professionals found that  77% agree the gap is widening between firms where AI is embedded in daily workflows and firms that use it only occasionally. Earlier this year, accounting practice management platform Karbon found 31% of respondents were concerned that the gap between AI-positive and traditional firms will continue to widen, a 4 percentage point increase from the previous year. 

Data from Thomson Reuters, though, implies this divide might not last and might be better described as a compression of the moderate-use middle with a persistent nonadopter tail. According to the company's 2025 and 2026 surveys for its AI in Professional Services Report, the proportion of firms saying they had no plans to use generative AI dropped from 28% to 19%, while those saying they were already using the technology grew from 22% to 40%. Another Thomson Reuters report suggests that the real gap is not between those who do and do not use AI — as the vast majority of professionals use the technology at least once a week if not daily — but between those who do and do not get value from their AI efforts: Over 90% of professionals are experiencing some degree of "AI value gap" where AI is apparently delivering for their organization but not for them personally. 

This speaks to a point raised by the authors of the AI-Native Firm Advantage Report: Their findings do not suggest a crude "more AI = better than" framework, but those who have shaped their firms around the technology versus using it to enact current processes more efficiently will be the ones best positioned to capture its value. A spokesperson for the study's authors noted that Active Deployers didn't just buy more AI, but 87.3% of them formally redesigned their organizational structures — including roles, team structures and staffing ratios — to fit an AI-enabled future. It's all in the operating model. Rolling out AI as a layer on top of how a firm already works is just a software purchase, while treating it as a reason to rethink how the firm works is a transformation decision.