AI isn’t coming to accounting, it’s already here. From auto-classifying transactions to enhancing audit procedures, the technology is embedding itself across the profession. But firm-level adoption remains cautious.
Why? Because the biggest barrier isn’t tools. According to Pascal Finette, an AI futurist and collaborator on CPA.com’s ongoing AI initiative, the most significant barrier to adoption is trust.
In a recent DCPA webinar, CPA.com EVP Michael Cerami and Pascal discussed what’s holding firms back from maximizing the potential of AI.
Firms are still figuring AI out
Rather than rushing adoption, leading firms are taking time to understand where AI makes sense in their workflows, how to validate its output and how to train staff to interact with it effectively. This measured approach aligns with the profession’s commitment to integrity, objectivity and competence.
More than half of firms are still in the experimentation phase when it comes to AI. According to Pascal, another 20% report only moderate use. And it's not surprising, because 95% of generative AI (GenAI) pilots fail, according to MIT’s State of AI in Business 2025 study. This mirrors what we see across the profession: lots of curiosity and growing capability, but still a very real confidence gap.
That gap is understandable — the pace of change has been extraordinary. The tools and use cases that firms are navigating today were seemingly implausible just five years ago. They are taking a deliberate approach to ensure the proper guardrails, policies and education are in place before scaling.
Confidence will unlock AI adoption
Pascal noted that firms aren’t holding back because AI tools aren’t ready. They’re holding back because they don’t fully trust the tools or how to govern them.
This is why the concept of "human in the loop" is so important. AI is not eliminating the need for professionals, but it is elevating them into critical new supervisory — and advisory — roles. For example, when firms use AI to auto-categorize financial data, an experienced professional must still validate and sign off on the report. The powerful combination of AI + human judgement ensures that output is reliable, explainable and secure. That’s when trust is earned.
Understanding this dynamic helps firms focus less on chasing every new tool and capability and more on designing a governance model that integrates AI safely and in the right way to create more value for the firm and their clients over time.
Rethinking the talent pipeline in an AI-enabled firm
As AI automates traditional entry-level tasks, a new question arises: How do early career professionals gain the insight and judgment needed to be advisors?
Historically, junior staff learned by doing basic processes, data entry and documentation. That model is being disrupted, and firms must rethink how they expose early-career talent to higher-order work and upskill them more quickly.
Firms of all sizes are reevaluating how they onboard, train and mentor new talent. Rather than relying on years of transactional work to build professional maturity, they need to create intentional learning experiences that fast-track judgment, strategic thinking and business acumen.
While entry-level professionals may arrive with more digital fluency and curiosity than ever before, firms will be challenged to channel that potential into roles that will grow their human expertise — not just manage AI outputs.
AI-forward firms will create a competitive advantage
Efficiency may be the initial driver of AI interest, but differentiation will be the long game. AI-native firms that embed intelligence throughout their workflows and decision-making will increasingly outpace their peers.
That doesn’t mean every firm needs to build proprietary models or in-house AI labs. In fact, most shouldn’t. But it does mean asking the right questions of your tech partners and evaluating solutions not just for automation, but for transparency, explainability and alignment with your service model.
Today’s AI is the worst it will ever be. It will only get better. That’s why firms shouldn’t wait for perfect systems — they should start building the infrastructure of trust now. Confidence will come from strategic experimentation, clear governance and continuous learning.
Because at the end of the day, the firms that lead won’t be the ones that moved the fastest. They’ll be the ones that moved the smartest.
To learn more about AI in accounting, explore CPA.com’s AI initiative page and our latest resources:
