CPA Firms Face Behavioral Hurdles in AI Adoption Despite Rising Use

Accounting firms understand that artificial intelligence represents a strategic opportunity. The challenge lies in translating that conviction into consistent practice across tax, audit, and advisory operations.

FloQast's 2026 research reveals the scope of the problem: 85% of accounting teams identify AI as a strategic priority, yet only 10% deploy it extensively. Just 17% of teams report feeling prepared to leverage their AI investments effectively. Meanwhile, separate 2026 research from CPA.com and Blue J surveying more than 1,000 tax professionals found that 60% now use AI-powered tax research at least weekly, up from 33% the previous year. That jump demonstrates what happens when tools address specific problems and professionals understand their application—but scaling from isolated successes to enterprise-wide adoption remains difficult.

Why it matters

The gap between strategic intent and operational reality in AI adoption represents more than a training deficit. It signals unresolved concerns about job security, professional identity, and unclear governance that no amount of technical demonstration will overcome. Firms that address the human dimensions of AI implementation—trust, role clarity, and psychological safety—will scale adoption faster than competitors focused solely on technology deployment.

Three behavioral barriers blocking firmwide adoption

Firm leaders often interpret uneven AI adoption as a knowledge problem requiring more training sessions or prompt-writing workshops. Frequently, however, employees have already formed emotional judgments about what AI means for their careers and professional standing.

The first resistance pattern centers on economic displacement. When employees believe that improving AI proficiency helps leadership eliminate positions, enthusiasm from executives sounds threatening rather than motivating. RSM's 2026 Middle Market AI Survey found that 85% of respondents reported executive leadership showing more AI enthusiasm than employees—a gap that transforms adoption into a trust issue before it becomes a skills challenge.

Where credible, firms should commit that productivity gains will first support growth, enhanced client service, and higher-value work rather than automatic headcount reduction. Making the career logic explicit helps: professionals who learn to supervise AI, validate outputs, and apply it responsibly become more valuable as workflows evolve.

The second pattern involves threats to professional identity. Accountants have invested years developing expertise in research, analysis, and judgment. When a tool performs that work in seconds, telling professionals AI will boost efficiency can miss the resistance source. They may interpret the message as their competence-proving work mattering less.

The more effective response involves professionals in workflow redesign. Ask which process elements feel repetitive and low-value, which require judgment, and where clients benefit most from human involvement. Starting with focused use cases, clarifying oversight requirements, and connecting AI learning to real engagement scenarios helps professionals see AI removing friction while their expertise shifts toward exceptions, recommendations, and client conversations.

The third pattern emerges when employees see useful AI applications but lack clarity about what the firm permits. They may experiment quietly, conceal how they produced drafts, or use unapproved tools because approved processes feel slower than the work itself. Research reported by CPA Practice Advisor found that one-third of lawyers, accountants, and compliance professionals were using AI their organizations had not approved, with rates rising among those who believed their organization moved too slowly on AI policy.

Making responsible use the default path

Firms need rules employees can apply during ordinary workdays. Specify approved tools, define which client and firm data can enter them, clarify when outputs require source verification or second-person review, explain what work should never be delegated to AI without human judgment, and create escalation paths for uncertain cases. This governance framework enables teams to expand AI use while preserving visibility, accountability, and review.

Professional obligations already point this direction. Circular 230 establishes competence, diligence, and conduct standards for tax professionals practicing before the IRS. Workable AI policies should translate those duties into concrete behaviors including verification, confidentiality, documentation, and supervision.

Psychological safety matters because people need permission to surface uncertainty. An employee should be able to say they used an approved tool, the result looks plausible, and they're unsure it's correct—without fearing that responsible experimentation will be treated as incompetence.

Peer champions and behavior measurement

Firmwide adoption won't spread through policy documents alone. People watch colleagues they trust. Choose peer champions for credibility rather than enthusiasm—professionals who understand the work, know where AI fails, and will show both useful applications and mistakes.

Measurement should focus on behavior change rather than license counts or login statistics. Thomson Reuters' 2026 AI in Professional Services Report found that only 18% of professionals said their organizations track AI return on investment, while another 40% didn't know whether ROI was measured. Better metrics track whether approved tools appear in selected workflows, whether employees verify outputs as required, how cycle time changes, and where saved capacity goes.

The details were first reported by CPA Practice Advisor, drawing on research from behavioral scientist Gleb Tsipursky and multiple industry surveys documenting the adoption gap between AI investment and operational integration.