AI won't fix your books.It'll just get you to the wrong answer faster.

Every owner I talk to is being sold AI right now. Your ERP vendor has an AI add-on. Your bank is using it for loan spreads. Someone on LinkedIn says it will replace your accounting team by next spring.

I use AI every day, and I'll tell you how in a minute. But here's the part nobody selling it wants to say out loud: AI is only as good as the business it's reading. Point it at messy books and it doesn't clean them up. It reads the mess faster, with more confidence, and hands you a polished answer that's wrong.

Most companies aren't ready, and it isn't the AI's fault

Here's what I see when I get inside a business that wants to "do something with AI":

• The same supplier set up three times in the vendor list, under three slightly different names.
• A chart of accounts built ten years ago for the tax return, not for running the business.
• Balance sheet accounts nobody has reconciled in months, because the person who knew how left.
• Margins that look healthy because something unrelated is propping them up.

None of that is new. AI didn't create those gaps. It just removed the excuse for leaving them open.

What a model would have missed

A couple of real examples, with the names left out.

At a manufacturer, $250,000 in rebates owed to the company was sitting on the balance sheet, accrued and never collected. Every report said the money was there. An AI tool reading those reports would have said the same thing. It took someone asking "has anyone actually been paid this?" to find it.

At a $130M fertilizer company, the margins looked strong. They were being propped up by credits that weren't coming back. Feed those numbers into a forecasting model and you get a confident growth plan built on a margin that didn't exist.

In both cases the problem wasn't a lack of technology. It was a lack of someone who knows what the numbers are supposed to look like, asking the uncomfortable question.

AI changes the cost, not the trust

This is how I think about it. AI makes the work cheaper and faster. It does not make the work trustworthy. Your banker doesn't trust a model. Your buyer doesn't trust a model. They trust the person who stands behind the numbers.

So AI is why I can do in days what used to take a team weeks, at a price a $10M business can afford. It is not the reason anyone should believe the answer. That part is still on me.

How I actually use it with clients

1. To build in front of you. A cash flow model, a pricing tool, a dashboard, a CRM your team will actually use. Things that used to be a six-week software project, I can often stand up in the room with you and shape while we talk.
2. To do the grinding faster. Matching, sorting, finding the duplicates and the outliers. The work that used to eat the first two weeks of every engagement.
3. Never as the final word. Every number that leaves my hands has been checked by me. If I can't explain it, it doesn't go in front of you, your bank or your board.

Before you buy an AI tool, do these three things

1. Clean the master lists. One record per customer, one per vendor, one per product. This is boring, and it's the highest-return hour you'll spend this year.
2. Make the chart of accounts match how you manage. If you run the business by crew, location or product line, your books should be able to tell you margin that way.
3. Give the data an owner. Not IT, not "accounting." A name. When nobody owns it, finance blames IT, IT blames finance, and the duplicate vendors stay.

Do those three things and AI becomes a real advantage. Skip them and you've bought a faster way to be wrong.

If you're not sure where your business stands, that's exactly the conversation I like having. Give me one problem in your business, and I'll show you what we can do about it.