A message to bookkeepers and accountants

When open-book exams were first introduced, people assumed they’d be easy. The formulas were right there. The textbook was open. In many cases, students even had access to previous tests.

And yet, many still struggled — because it turns out, knowing where the answer lives isn’t the same as knowing how to use it.

That’s what’s happening right now with AI. And if you’re a bookkeeper, tax preparer, or accountant watching the headlines, I imagine it feels personal.

So let me be straightforward: AI is going to change your job. It’s already changing it. But it’s not going to replace the thing that actually makes you good at it.

A recent MIT Media Lab study found that when people relied on AI to complete their work, their neural connectivity — the brain’s way of forming and reinforcing understanding — didn’t just dip while using the tool. It stayed dampened even after the AI was taken away. The thinking had started to atrophy.

In other words: the people who handed their thinking over to AI got worse at thinking. Not while they were using it — after.

That finding should worry some people. But it should reassure you. Because it means the gap between someone who understands the work and someone who can operate a tool is getting wider. And you are on the right side of that gap.

The reason AI can’t replace a good bookkeeper or accountant isn’t that the technology isn’t smart enough. It’s that the job isn’t what most people think it is. People outside this profession think accounting is about calculations. You know it’s about judgment. It’s knowing that when Q4 profitability “improves” in a report, you check whether payroll got misclassified or a large AP entry didn’t hit. It’s knowing that a 25% drop in marketing spend might just mean the Stripe feed cut off on March 28th. AI won’t flag a missing import. You will.

That kind of knowledge doesn’t come from a formula or a prompt. It comes from years of seeing things go wrong and learning what “wrong” looks like before the numbers even confirm it.

My dad understood this instinctively. He didn’t let us use calculators growing up. Not because he was anti-tech — it wasn’t some principled stance against progress. It was practical. He wanted us to be fluent in the work. Not just capable when everything was working perfectly, but confident even when the batteries died, the app crashed, or the Wi-Fi dropped.

That’s the same thinking behind how we’re rolling out AI tools at Numera. We’re investing in AI across our teams — tools that make our people faster, help them surface patterns, and catch things that would’ve taken an hour to find manually. But every time we deploy a new tool, we tell our team the same thing: the tool is the copilot. You are the pilot. And we need you sharp — because the day the tool gets something wrong, and it will, we need you to be the person who catches it.

Here’s what staying sharp looks like in practice:

Check the AI’s math against reality, not just logic. If your AI tool tells you something changed, go check why. Did the client actually change something, or did a data feed break? AI sees the numbers. You see the story behind them.

Remember what’s missing. AI only works with what’s in front of it. You carry context it doesn’t have — last quarter’s conversations with the client, the invoice that always comes in late, the pattern you noticed three months ago. That institutional memory is yours. No model has it.

Reconcile something by hand once a week. Pick one bank rec or one trial balance and walk through it line by line without AI assistance. Not because the tool would’ve gotten it wrong — but because the day it does, you need to be the person who notices.

Fix bad prompts, not just bad outputs. If your AI tool gives you a jumbled answer, don’t just re-prompt. Ask yourself whether you framed the question clearly. “Summarize this P&L” is too vague. “Compare COGS and overhead trends over the past 3 months and flag anything that doesn’t align with cash flow movement” — that’s a question worth answering. The tool will only be as sharp as you are.

I know the headlines make it sound like AI is coming for your seat. What I see is the opposite. The tools are getting smarter, and that’s making the distance between someone who understands the work and someone who’s just pushing buttons more obvious than ever.

You understand the work. That’s not something anyone can automate. It’s something worth continuing to build.

I know the headlines make it sound like AI is coming for your seat. What I see is the opposite. The tools are getting smarter, and that’s making the distance between someone who understands the work and someone who’s just pushing buttons more obvious than ever. Your clients aren’t losing sleep over which AI tool their accountant uses. They’re losing sleep over payroll, over cash flow, over whether they’ll make it through the next quarter. They need someone who gets it — not a tool that processes it. That someone is you. It always has been. AI just made it easier to see.