How to Actually Make AI Work in Your Accounting Firm

Most firms have a system problem that AI is exposing

Everyone’s rushing to experiment with AI right now. Well, at least all the firms that will matter in five years are.

They’re adding new tools, testing automations and sitting through demos that promise to make everything faster and more efficient. And for a moment, it feels like progress.

But then… not much changes. The workflows are still messy. The team is still overwhelmed. The impact feels smaller than expected.

So what’s actually going wrong?

It’s not that AI doesn’t work. It’s that most firms aren’t set up for it to do anything helpful.

AI just amplifies broken systems. And accountants’ systems have been broken for a long time.

AI is only as good as your workflow

There’s a common misconception right now that adding AI will naturally make things more efficient. But AI doesn’t replace your systems, it runs on them.

It pulls from your data, your workflows and your processes. If those things are inconsistent or unclear, AI doesn’t clean them up — it scales them. That’s why some firms see real gains, while others feel like they’re spinning their wheels.

The difference isn’t the tool. It’s the system underneath it.

AI isn’t breaking your firm, it’s revealing it

Most of the time, when AI “fails,” it’s exposing problems that were already there.

Data is scattered across emails, spreadsheets and different tools, with no real source of truth. Workflows vary depending on who’s handling the task, making it hard to create consistency. Leaders don’t always have visibility into where work slows down or breaks, which makes it even harder to improve anything intentionally.

And underneath all of that, many processes were never actually designed. Things just piled on over time.

That kind of environment makes it nearly impossible for AI to do its job well.

Fix the system, not the tool

If firms want to see real impact from AI, the focus has to shift. Not toward more tools, but toward better systems.

That means getting data organized and accessible. Defining how work should flow, even if it’s not perfect yet. Creating visibility into what’s actually happening across the firm, instead of relying on assumptions.

Once that foundation exists, AI becomes powerful. Without it, it’s just another layer of complexity.

So the real question isn’t, “Which tool should we use?”

It’s, “What kind of system are we giving it to work with?”