Pevita Pearce
CPA firms are acutely aware of artificial intelligence’s transformative potential, yet the practicalities of implementation remain a significant hurdle. Research consistently highlights a stark contrast between strategic intent and actual adoption. A 2026 study by FloQast revealed that while 85% of accounting teams consider AI a strategic priority, only a modest 10% are utilizing it extensively. Furthermore, a mere 17% of these teams feel adequately prepared to leverage their growing investments in AI. This considerable gap underscores the critical need for an AI adoption playbook that prioritizes human behavior with the same rigor as technological integration.
The technological integration of AI into accounting workflows is already well underway, demonstrating its capacity to solve specific challenges. A 2026 survey conducted by CPA.com and Blue J, encompassing over 1,000 tax professionals, indicated a significant surge in AI adoption for tax research. The study found that 60% of respondents now employ AI-powered tax research at least weekly, a substantial increase from 33% the previous year. This rapid adoption trajectory in accounting AI, when a tool directly addresses a professional need and its utility is clearly understood, offers a glimpse into successful implementation. The persistent challenge lies in scaling these localized successes to achieve consistent integration across the diverse functions of tax, audit, advisory, and overall firm operations. Emerging research into the middle market reveals that even successful AI pilot programs frequently encounter scaling barriers. These obstacles often include issues related to data quality, system integration complexities, cybersecurity concerns, and crucially, the readiness of the workforce to adapt.
Diagnosing Resistance: Beyond the Training Session
A common misstep by firm leaders is assuming that uneven AI adoption stems solely from a lack of technical proficiency, leading to an overemphasis on additional demonstrations or prompt-writing workshops. While such training can be beneficial, the root of resistance often lies deeper, rooted in employees’ emotional responses and their perception of AI’s impact on their professional lives.
One significant driver of apprehension is the fear of economic displacement. If employees perceive AI as a tool that could automate their roles and lead to job losses, leadership’s enthusiasm can be misconstrued as a threat rather than an opportunity. The 2026 RSM Middle Market AI Survey corroborated this sentiment, reporting that 85% of respondents indicated a notable gap between executive leadership’s optimism regarding AI and employee confidence. This disparity highlights that firmwide AI adoption is, at its core, a matter of trust before it becomes a skills deficit.
To bridge this trust deficit, firms should proactively articulate a job-neutral commitment, emphasizing how AI-driven productivity gains will primarily support business growth, enhance client service, facilitate higher-value work, and enable redeployment of talent, rather than immediately targeting headcount reduction. Making the career logic explicit is also crucial: professionals who develop the skills to supervise AI, critically evaluate its outputs, and utilize it responsibly will become increasingly valuable as workflows evolve. This reframes AI not as a job eliminator, but as a catalyst for professional development and enhanced career prospects.
Another layer of resistance emerges when AI is perceived as a threat to professional identity. Accountants dedicate years to honing expertise in areas such as research, analysis, writing, review, and critical judgment. When AI tools can perform aspects of these tasks with unprecedented speed, messages of increased efficiency can inadvertently overlook the core of employee resistance. Professionals may interpret these messages as devaluing the very work that validates their competence. This reaction warrants careful consideration, especially as accounting leaders increasingly view AI as a means to offload routine tasks, thereby freeing up professionals for higher-value activities like strategic planning, relationship building, and cultivating client trust.
A more effective approach involves actively involving employees in the redesign of workflows. Engaging teams in discussions about which processes feel repetitive and low-value, which require nuanced judgment, and where human interaction provides the most significant client benefit can foster a sense of ownership and collaboration. Publications like CPA Practice Advisor have consistently advocated for starting with focused use cases, clarifying oversight responsibilities, and directly linking AI training to real-world engagement scenarios. These practical AI guardrails help professionals visualize a future where AI streamlines administrative burdens, allowing their expertise to be channeled towards complex problem-solving, strategic recommendations, and more meaningful client interactions.
Enabling Responsible Use: Making it Easier Than Going Rogue
A third common behavioral pattern emerges when employees recognize the utility of AI applications but are unclear about the firm’s official stance or permitted usage. This ambiguity can lead to "shadow AI," where individuals experiment with AI tools covertly, conceal their AI-generated work, or adopt unapproved tools because the sanctioned processes appear cumbersome or slow. CPA Practice Advisor reported that a significant one-third of legal, accounting, and compliance professionals were utilizing AI tools that their organizations had not officially sanctioned, with this rate escalating among those who felt their organizations were lagging in AI implementation. This prevalence of shadow AI is a predictable outcome of unclear policies and a lack of accessible guidance.
To counteract this, firms must establish clear, actionable guidelines that employees can easily apply in their daily tasks. This includes explicitly naming approved AI tools, defining the types of client and firm data that can be input into these systems, and specifying when outputs require source verification or a second review. It is also imperative to delineate tasks that should never be fully delegated to AI without human oversight and judgment. Establishing a clear escalation path for uncertain scenarios provides a safety net for employees navigating new AI applications. This form of governance control empowers teams to expand their use of AI while maintaining crucial visibility, accountability, and robust review processes.
Existing professional obligations provide a foundational framework for responsible AI use. For instance, Circular 230 establishes stringent standards for competence, diligence, and conduct for tax professionals practicing before the IRS. A well-designed responsible AI use policy should translate these overarching duties into concrete AI-related behaviors, encompassing verification of AI outputs, maintaining client confidentiality, thorough documentation, and diligent supervision of AI-assisted work.
The concept of psychological safety is paramount in this context, as it empowers individuals to voice uncertainty without fear of reprisal. Groundbreaking research by Amy Edmondson highlights the critical link between psychological safety and learning behaviors within work teams, fostering an environment where open communication and the discussion of errors are encouraged – precisely what is needed during AI experimentation. An employee should feel comfortable stating, "I used the approved AI tool, and while the result appears plausible, I’m uncertain about its accuracy," without the risk of their responsible experimentation being misconstrued as incompetence. Such openness provides managers with invaluable opportunities to guide and reinforce judgment before potentially flawed AI outputs reach clients.
Leveraging Peers: Catalyzing Behavior Change
Achieving firmwide AI adoption extends beyond the dissemination of policy documents. Professional behavior is significantly influenced by observing trusted colleagues. A skeptical audit manager, for example, is far more likely to reconsider their stance on AI after witnessing a respected peer effectively integrate it into a familiar workflow than after attending a generic vendor demonstration.
The selection of peer champions should prioritize credibility and a deep understanding of the work, rather than mere enthusiasm. These individuals should be adept at identifying where AI excels and where it falls short, and be willing to transparently showcase both successful applications and any encountered mistakes. Their role is to model effective AI habits within the actual tax, audit, advisory, and administrative workflows of the firm. Furthermore, they serve as a vital conduit for feeding recurring challenges and insights back to leadership. The accounting profession already offers examples of firms that have successfully paired enterprise-wide AI deployment with comprehensive enablement strategies, including role-specific training paths and explicit, actionable guardrails. CPA Practice Advisor has reported that existing skills shortages are already impeding AI and automation initiatives, underscoring that practical, role-based AI training is as much a capacity-building exercise as it is a learning endeavor.
Managers play a crucial role in reinforcing desired behaviors. Integrating AI-related discussions into performance reviews – inquiring about where AI proved beneficial, where it faltered, what was verified, and what adjustments are needed for future use – transforms AI engagement from an ad-hoc experiment into an intrinsic part of the firm’s learning and development process.
Measuring Progress: Beyond License Counts
Assessing AI adoption solely by license counts, login frequencies, or training attendance provides leaders with data on access and engagement, but it fails to reveal whether actual work processes have transformed. The 2026 Thomson Reuters AI in Professional Services Report indicated that only 18% of professionals reported that their organizations actively track AI return on investment, with an additional 40% unsure about current ROI measurement practices. A more meaningful approach to measurement begins with identifying the specific AI behaviors that firms aim to cultivate.
Tracking the integration of approved tools into selected workflows, the diligence with which employees verify AI outputs, the trend in review corrections (whether rising or falling), changes in cycle times, and the allocation of freed-up capacity offers a more nuanced understanding of AI’s impact. Supplementing this with concise pulse surveys can gauge employee confidence: Do professionals know which tools are permissible? Are they aware of data restrictions? Do they feel secure admitting when an AI-assisted draft requires further refinement?
The ultimate objective is to establish a system where experimentation is normalized, errors are identified early, and effective practices are disseminated throughout the firm. This emphasis on measurable implementation is increasingly recognized within the accounting industry, with awards and acknowledgments being given to firms that effectively combine AI deployment with robust governance and replicable outcomes. CPA Practice Advisor has articulated that fostering an AI-first culture hinges on the development of everyday habits and diligent human review, rather than solely on the acquisition of new technologies. This focus on AI integration into daily operations is the critical differentiator between isolated productivity gains and a firm that fundamentally transforms its operational paradigms.
CPA firms face strong imperatives to accelerate their AI adoption. However, a rapid deployment without careful consideration of behavioral dynamics can result in a fragmented landscape: early adopters surge ahead, hesitant professionals lag behind, and discreet users develop their own informal protocols. A pragmatic approach to AI adoption in the workplace directly addresses these varied behaviors. It involves providing employees with credible reasons to embrace change, preserving the elements of professional identity that are most valued, making governance guardrails user-friendly, empowering trusted peers to demonstrate best practices, and diligently measuring whether behaviors and business outcomes evolve in tandem. This comprehensive strategy is the pathway for AI to transition from sporadic experiments to a pervasive and impactful firmwide capability.
