Editor’s Note: This article is part 2 of a 2-part series on AI adoption in accounting firms. View all of the articles in this series here: AI Adoption in Accounting Firms
In Part 1, we talked about why AI adoption in accounting firms stalls. It is not because accountants do not care. It is not because the technology is not interesting. It stalls because firms are wired to avoid mistakes. AI feels risky in a profession where mistakes are expensive.
That fear is real. It is also expensive.
While leadership is still discussing AI, the team is using it. They are cleaning up emails, summarizing meetings, organizing notes, drafting internal memos, and fixing clunky writing. They are turning rough thoughts into something another human can actually read without developing a twitch. The question is not whether your people are using AI. The question is whether your firm is going to lead that learning. Or will you keep pretending it is not happening?
Your firm needs to move from talk to action. Not with an innovation summit. Not with a fifty-page policy manual nobody reads. And definitely not with a software shopping spree dressed up as strategy. Buying a tool does not build a capability. If it did, joining Weight Watchers would automatically make you thin.
Too many firms treat AI as a technology decision. It is not. It is a leadership decision. An operating decision. A behavioral decision. Firms making progress are not the ones with the longest list of subscriptions. They are the ones creating permission, boundaries and shared learning. That moves firms forward.
Let’s make it your firm.
The first step is permission. Most firms send mixed messages. Leadership says AI matters. Then nobody tells the team where it is okay to test it, what is off limits, or what still requires human review. People do what people always do in unclear environments. They freeze or work around the uncertainty on their own. Neither one helps your firm.
This is where firms get a little ridiculous. They say they want innovation. Then react to experimentation like someone lit a fire in the break room. People stop sharing. Not because they are trying to be rebellious, but because curiosity is treated like a compliance violation.
You want movement? Say something clearly. Use AI for low-risk internal work. Do not use it for confidential client information unless approved. Do not trust the first answer. A human reviews everything before it goes anywhere. That kind of direction lowers fear and raises accountability. People know where the line is. This matters because most firms do not have an AI adoption problem. They have a permission problem.
The second step is visibility. Too much AI learning happens in private. One person finds a smart use case. Another figures out a better way to write prompts. Somebody else learns where the output falls apart. None of this becomes firm knowledge. Nobody shares it consistently. The firm keeps relearning the same lessons one person at a time. That is not caution. That is waste.
You see this scene playing out in firms everywhere. Leadership forms the AI committee. The committee schedules the meeting. Then they hold another meeting to discuss what the first meeting surfaced. Everyone says the right things about governance, risk and responsible adoption. Meanwhile, two people on the admin team quietly figured out how to use AI to summarize notes, clean up communication and save an hour a day. Leaders are still discussing whether the swimming pool has enough chlorine. Guess what? The kids are already in the water.
Visible learning matters. You do not need an innovation lab. You need one place where people can share what they tried, what worked, what failed and what needs caution. A weekly meeting works. A shared document works. A Slack thread works.
The point is not glorifying the tool. The point is making learning visible. Once learning becomes visible, the firm gets smarter faster. Patterns emerge. Better practices develop. Judgment improves. AI stops being a random individual shortcut. It starts becoming a team capability.
The third step is to stop buying tools before defining the job. Buying feels productive. Demos feel productive. Vendor calls feel productive. Saying “We are exploring solutions” feels productive. But none of that matters if the firm has not answered the basic question: what problem are we trying to solve?
Before you buy the next tool, ask better questions. What work are we trying to improve? Where are we losing time on low-value effort? What keeps slowing the team down? What would make us more useful to clients?
That is the conversation that matters. Clients do not care that your firm added another piece of software. They care that you are easier to work with, faster to respond and better at helping them make decisions. If the technology does not improve the experience or the outcome, it is just overhead with a polished demo.
This matters. AI is not a race to automate everything. It is a way to remove friction, improve judgment and create more room for higher-value work. Better conversations, better follow-up, better analysis and better service. Used well, AI does not make your firm less human. It gives your people more capacity to be human where it counts.
That is the disruptive part. The old model says efficiency means doing more work with less effort. The better model says efficiency should create space for better thinking, better service and better client relationships. If AI only makes the machine run faster, that is not much of a win. If it helps your team communicate better, think better and serve better, now you are building something that matters.
So yes, talk about AI. But stop confusing talk with progress. Progress happens when leaders create permission. When teams share learning. When firms redesign work instead of chasing tools. That is how firms move from talk to action. That is how AI adoption in accounting firms becomes real.