AI has moved from an abstract concept to a daily reality inside CPA firms. The challenge is no longer whether firms should pay attention, but how they do so without creating risk, confusion or unrealistic expectations.
To get a grounded view of how CPA firms are actually approaching AI, we sat down with Gary Thomson and Kate Krupey, two advisors who spend their time helping firms navigate change without unnecessary disruption.
Gary Thomson is the founder of Thomson Consulting and a longtime advisor to managing partners across the country. Over a decades-long career, he has worked with firms of all sizes and brings a practical, partner-level perspective shaped by deep involvement in the profession, including leadership roles with CPA societies and the AICPA.
Kate Krupey is the VP of CPA Practice at Netgain and a former CPA firm CIO. She works closely with firms nationwide on technology strategy, change management and practical AI adoption, with a focus on making complex topics clear, useful and actionable.
Together, Gary and Kate see the same pattern playing out across the profession. AI adoption is accelerating, but structure and clarity often lag behind.
“Artificial intelligence isn’t a future conversation anymore,” Gary said. “It’s a present-day leadership issue.”
Adoption is moving fast. Maturity is not.
One of the biggest misconceptions Gary and Kate see firms encounter is being behind the AI curve. In reality, most firms are already using it, just not in a consistent or intentional way.
“What I’m seeing is very fast adoption of the first level,” Kate explained. “Firms are getting in the door quickly, but they’re struggling to move to the next stage.”
Staff are experimenting with tools. Vendors are embedding AI into tax, audit and workflow platforms. Clients are beginning to ask how firms use AI to improve accuracy and insight. What’s missing is a shared framework for how AI fits into firm operations.
“Experimentation is happening everywhere,” Kate said. “Governance and consistency are what slow firms down.”
Where firms are actually getting stuck
When AI initiatives stall, it’s rarely because the technology itself falls short. More often, the foundation is not ready.
“You can’t apply AI effectively if you don’t understand where your data lives or how your work actually flows,” Kate said. “AI amplifies what’s already standardized. If workflows aren’t documented or data is scattered, it’s very hard to get meaningful results.”
Gary sees the same challenges across firms of all sizes and markets.
“We’ve heard ‘this will change everything’ many times over the years,” he said. “AI feels different, but healthy skepticism is still part of the profession. The risk is letting skepticism turn into inaction.”
Firms that make progress tend to focus less on tools and more on process clarity, ownership and a small number of well-chosen pilots.
What’s working inside real firms
The most effective AI use cases are not flashy. They focus on removing friction from work people already do every day.
In tax, firms are using AI to summarize planning conversations into clear action items, draft memos with citations, and extract data from W-2s and K-1s to reduce manual review. In audit and assurance, teams are applying AI to risk brainstorming, anomaly detection and contract summarization, allowing them to spend less time assembling information and more time interpreting it.
“For me, AI just speeds everything up,” Kate said. “It gets me out of creation mode faster so I can spend more time thinking.”
That shift is critical. When teams spend less time drafting and compiling, they have more capacity for judgment and client-facing work.
“My favorite use case isn’t generation,” Kate added. “It’s analysis. Taking large amounts of unstructured data and getting insight quickly just wasn’t available to end users before.”
Some of the fastest early wins are happening outside of client service lines. Admin, HR and IT teams are using AI for email triage, self-service bots and internal documentation, reducing interruptions and improving consistency across the firm.
Demystifying bots, agents and AI tools
One common barrier Gary hears is the assumption that meaningful AI requires significant investment.
“When people hear ‘bot,’ they think it means a massive price tag,” he said. “What’s surprised me is how much firms can build internally with curious staff.”
Kate sees that curiosity as an advantage.
“Accountants actually have the right mindset for this,” she said. “They’re technical, process-driven and used to learning continuously.”
Firms that allow safe experimentation, supported by clear guidance, tend to uncover practical applications faster than those waiting for a perfect plan.
“If we’re not failing occasionally,” Gary said, “we probably aren’t exploring enough.”
Governance that enables progress
Governance is often viewed as the thing that slows innovation. Done well, it does the opposite.
“Governance shouldn’t slow you down,” Kate said. “It should give people confidence to use AI safely.”
Most firms have already taken an important first step by establishing an AI acceptable use policy. The next step is deciding how ideas are evaluated, approved and scaled.
“Good governance is about principles, not perfection,” Gary said. “That applies just as much to AI as it does to firm leadership.”
Clear expectations around approved tools, data boundaries and required human review reduce risk and eliminate guesswork for staff.
Rethinking ROI
One of the most common leadership questions is how to measure return on investment.
“ROI doesn’t really show up in the experimentation phase,” Kate said. “That stage is about learning.”
Early value often appears as reduced cycle time, improved consistency and higher morale, not immediate financial gains.
“You need a budget for failure,” she added. “Failure is learning. Repeating the same failure without learning is the problem.”
Gary frames it as a familiar progression for CPA firms.
“There’s a point where you invest to figure things out,” he said, “and a point where you start measuring returns. AI is no different.”
The mindset shift CPA firms need
Perhaps the biggest adjustment is cultural.
“This isn’t about perfection,” Kate said. “It’s about progress.”
AI rewards curiosity and iteration more than certainty. Firms that embrace a “what if” mindset tend to move faster and with more confidence than those waiting for complete clarity.
Why waiting is the bigger risk
Waiting is not neutral. AI is already part of how work gets done.
“Your people are already using AI,” Kate said. “The question is whether you’re giving them structure or leaving them to figure it out on their own.”
Gary sees this as a talent issue as much as a technology one.
“The next generation wants to work at firms that are curious and willing to adapt,” he said. “How we approach AI sends a strong signal about the future.”
Firms don’t need to chase every new tool. They do need clear guardrails, intentional pilots and a willingness to learn.
“AI isn’t something firms can afford to ignore,” Kate said. “With the right structure, it’s an opportunity to make firms better, not riskier.”
Want to go deeper?
You can watch the full on-demand conversation with Gary Thomson and Kate Krupey for additional examples, use cases and practical guidance.