At a recent industry event, my colleagues and I walked into a managing partner roundtable and asked the group to name their highest-priority topic. We were told to guess, and the answer was only two letters. In today’s landscape, we knew it was either “AI” or “PE.” I immediately said, “AI.” Sure enough, I was right.
The entire conference centered on those two letters—you couldn’t attend a session or walk the trade show floor full of technology vendors without hearing about it. No doubt, the conference made clear that the artificial intelligence (AI) era in accounting has arrived.
Yet, despite AI’s growing prevalence in our everyday work, work events, news, and extracurricular activities, I believe that whatever happens next is up to us. We can either succumb to the hype that accounting jobs will be hit hardest, or we can rise to the occasion and use AI to finally unlock the promise our profession has always had—the ability to be true advisors in everything we do.
So, how do we unlock that promise? Truth be told, it’s still early. Firms and their clients are still largely in AI pilot mode. But at the same time, the technology is moving faster than most of us can keep up with: OpenAI releases a new ChatGPT capability, then Claude jumps ahead, and before you can catch your breath, the conversation shifts again. But standing still also isn’t an option—you’ll fall behind. You also can’t go all-in everywhere at once because you’ll drown in investment dollars and change fatigue.
The answer? In my view, firms need a bottom-up, top-down strategy. Start with AI proficiency, then move on to strategy. And fair warning: Don’t start with a technology product (sorry, trade show vendors).
Importantly, AI can’t be layered on top of a firm’s tech stack or technology strategy that hasn’t done the foundational thinking. If the strategy is unclear, AI will only magnify that. Here’s a five-step roadmap for strategically unlocking AI’s promise.
We can’t expect to unlock the potential of AI without understanding the technology, its use cases, and the changing skills required to use it well. At the center of that is a skill we all need to get comfortable with: curiosity.
In many ways, learning AI is no different than how we learned as kids. We try something, it works or it doesn’t, and then we try again until we get better. The same is true with AI. Firms owe it to their people to provide them with the tools, training, and room to experiment and fail (of course, with guidelines and governance in place).
All of that doing, failing, learning, and doing again will produce something valuable: real use cases that create efficiencies, open new revenue opportunities, and help your people serve clients in a deeper way.
A governance framework isn’t something that happens organically. It needs to be thoughtful, intentional, and strategic. After all, the risks surrounding AI are real.
When creating an AI governance framework, firms should be asking:
Remember, if your rules are too tight, you’ll struggle to innovate. If they’re too loose, you open yourself up to unnecessary risk. The goal is to find balance.
Firms also need to protect clients’ personal protected information, establish clear expectations for how AI can and can’t be used, and decide how transparent they want to be with clients about that usage. And, of course, they need to think seriously about bias— where it may show up and how they’ll guard against it.
Once governance is in motion, firms need to inventory the workflows and jobs being done across the business. This is where the use cases come into focus.
The easiest use cases to identify are productivity use cases. These are the opportunities to free up hours by removing friction and making routine work more efficient. They exist across service lines and within the operational functions that support the firm. To find them, firms need to understand the workflows happening every day and where the pain points are located.
From there, the goal is to document the use cases, assess how meaningful they would be if solved, and weigh that against the ease of implementation. Just as important, firms need to understand the return on investment. For example, if the change works, how many hours are you giving back to the team? This, of course, then leads to an important follow-up question: What will you do with those extra hours?
Remember, efficiency gains are great, but they only matter if you deploy the capacity well. Will those extra hours go toward better insights? Greater client service? Serving more clients? Launching new services? Rethinking the talent model? Those downstream decisions matter just as much as the initial gain.
Productivity gains are the easiest to get our arms around, but they may not be the most impactful over time. Some of the most powerful use cases will come from doing things differently and creating new value.
Identifying and implementing opportunity use cases will require more creativity, better data structure, and more planning, but they also hold significant upside. To do so, ask:
There’s another category as well for firms to consider—imaginative use cases. These are often bottom-up in nature and practitioner led. They allow professionals to serve clients in a deeper and more complete way by extending their own expertise. When done right, AI has the potential to amplify human judgment, not replace it.
That distinction matters. AI can’t become a crutch—it shouldn’t do our thinking for us. But it can absolutely extend our thinking, sharpen it, and help us apply our expertise more broadly and more effectively.
Now the vendors get their turn. Software providers hold a lot of promise for the future of AI in the profession. For many firms, these technologies will create access to capabilities that, in the past, were largely available only to the largest firms. In that sense, AI has the potential to democratize sophisticated tools and allow boutique firms to compete in new ways. But there’s a catch. This is where use case prioritization matters.
If we implement technology for technology’s sake, we’ll spend a lot of time and money without the return to justify it. There will absolutely be strong cases to partner with technology platforms. There will also likely be a place to internally build an AI platform, especially in the orchestration layer where firms connect workflows, systems, and data in ways that fit how they actually operate.
The point is not to buy or build on principle. The point is to make that decision based on the use case, economics, and strategic value to the firm.
There’s a tendency in moments like this to want to skip ahead and jump straight to transformation. But we can’t jump straight to that stage. It’s likely still a few years out. But that doesn’t make the work any less urgent. If anything, it makes the next steps more important.
The potential of AI will present us with many opportunities, and if we approach it with the right balance of curiosity, discipline, strategy, and imagination, it can finally unlock what this profession has always been capable of becoming. For firms willing to do the work, AI may prove to be one of the greatest opportunities our profession has ever had.
