Episode 71: AI Bookkeeping Tools: The Hidden Risks Creative Entrepreneurs Need to Know

Welcome to the creative mind smart money podcast, where we turn financial confusion into creative confidence. I’m Samantha Eck, bookkeeper and fractional CFO for creative entrepreneurs. Each week, I’m sharing my financial expertise and actionable strategies to help you build a thriving creative business.
Plus, you’ll hear from industry experts who bring fresh perspectives on growing your business beyond the numbers. Because building a successful creative business starts with strong financial foundations. Your next chapter starts now.
You’re listening to creative minds, smart money podcasts. And today’s topic is on the dangers of AI in bookkeeping. Okay.
And you might be like, wow, Samantha, that’s a really bold topic because AI is everywhere nowadays. And AI is exciting and new and fresh. And trust me, I know I love AI.
I am utilizing it as a tool to its fullest extent that I believe I can. But I want you to remember that word tool. AI is a tool.
It is not the end-all be-all. It cannot replace a human being yet. It cannot replace relationships.
And it cannot replace, you know, quality control. There is so much that AI is not yet capable of. Obviously, in coming years, it may be.
But I just want you to remember that there are human beings out there as well. And there’s an environmental impact using AI as well. So, there is a lot of different factors that go into AI and things like that that we really just need to think about and be aware of.
Okay? So, first of all, there’s AI tools everywhere promising to automate bookkeeping and make life easier. Maybe they’re automating, you know, the categorization. Maybe they’re automating the reconciliation.
Maybe they’re automating everything out there. There’s a lot of temptation because obviously, who doesn’t want their books handled instantly and as cheaply as possible? Of course, as business owners, yes, we want to save money. But there are some things that you don’t necessarily want to be saving money on.
Like, there’s some things that you really want to invest in. And I promise you that the financial aspect of your business with like bookkeeping and things like that is something you want to invest in. It is not something that you just want to toss out there and say, hey, like, just take it away from me.
I don’t want anything to do with it. You want to invest in it. Okay? So, yes, AI is powerful.
But it is not a substitute for the context of your business. It’s not a substitute for the judgment of your business. And it’s not a substitute for the strategy of your business.
Okay? So, I want to remind you and think about those things. Now, where does AI show up in bookkeeping currently? Like, as we’re looking at this whole bigger picture, where is it showing up? So, right now, it’s an automated transaction categorization in tools like QuickBooks, Zero Wave. Zero doesn’t have it as much.
QuickBooks just overhauled their whole system software. And it’s huge. Like, the categorization, the transaction categorization thing is huge.
They also have, you know, reconciliations. They have a bunch of stuff that’s going on. And I have tested each and every one of it.
And I will tell you that the AI categorization is almost worse than what they used to have when they tried to just suggest a category without AI. So, it’s just… it’s not worth it right now. But we’re going to get into it.
We’re going to talk about the risks in a second. Okay? And then obviously a big feed rules and machine learning that kind of predicts where expenses should go. And some of those are great, honestly.
Some of those are fantastic. But there’s a risk in that as well. Again, we’re going to get to that.
We’re going to talk about that. There are, of course, just AI bookkeeping apps in general that are marketed as bookkeeping without a bookkeeper. So, you don’t need a bookkeeper at all.
You just need bookkeeping. Now, one of the softwares that I want to talk about is Puzzle. There is a software out there called Puzzle.io. It advertises AI in bookkeeping, but it also advertises the importance of also having a bookkeeper.
So, it’ll do the grunt work for you, but then it’s like you still need a bookkeeper to double check things. You still need a bookkeeper to do your software. So, they’re a company that believes in AI, but also believes in the person that should also be there to kind of oversee everything.
So, this is where the understanding of what a bookkeeper is is slowly starting to evolve and change. Yes, we can categorize things. Yes, we can do all this stuff.
But a bookkeeper is someone who’s deeper than that, right? I’ve told you before, a bookkeeper is someone who helps you strategize. A bookkeeper is someone who helps you analyze your whole business. But we can’t do that if we’re not seeing the data, right? So, if you already have that data, then that’s where we can just become more of a strategic partner and less of that someone who’s in there doing the data entry.
And obviously, that strategic partnership will come with a higher cost, but that’s why, you know, you have something that’s cheap that gets your bookkeeping done. And then when you really need that help to zero in on all the strategy you have. Anyways, getting off topic, let’s talk about the real risks of using AI in your business.
So, first of all is misclassification. AI does not understand nuance. It doesn’t.
And it never will. So, it’s not going to understand what a client gift is. So, if you go to Starbucks and you bought a gift, like a Starbucks gift card versus a meal.
So, again, if you go to Starbucks and maybe you bought a coffee and marketing. Again, if you go to Starbucks and maybe they’re offering sponsorship spots at the front of their window, you could like put your logo or something. It’s not going to differentiate between that.
It doesn’t understand the nuance. It sees Starbucks and it’s like, oh, Starbucks is a coffee company. It’s automatically meals.
So, those wrong categories can really distort your profit margins and then also distort your tax deductions. And then when you get audited, that can cause a lot of issues. So, I just want to be clear on that.
Like, the IRS is not something you want to mess around with. And there are so many people who will just go with whatever like the AI is telling you. And that’s not always the right move.
You want to make sure it’s in the correct category. Anyone can categorize something. The robots aren’t there yet.
I’m telling you. I’m just being honest with you. AI will also have context blindness.
AI can’t ask why your revenue dipped, right? So, if I’m noticing that your revenue dipped, I’m going to say, hey, you know, Amanda, why is your revenue lower this month than last month? What was going on? And you could be like, hey, Samantha, you know, I went to less shows this week or last month. I just, I didn’t go to as many events as I thought I would. So, that’s why my sales were down.
Or I didn’t do as many networking events. So, that’s why my sales are down. When my sales are up historically, it’s because I’m going to a bunch of different events and doing a bunch of different things.
Okay? So, they also can’t differentiate whether a big expense was a one-time investment or not. So, if you have a really huge expense, maybe it’s just thinking that this is a software charge or something like that. It might not recognize it as an asset.
Maybe you bought a car and it’s like, okay, we’re just going to put that in, you know, somewhere on your expenses. It’s not going to be able to understand that. And it also won’t tell if payroll shifts are seasonal versus permanent.
So, maybe you have a heavy hiring period in the summer and then you slow down and kind of let some people go in the winter. It’s not going to understand that. It doesn’t understand the context behind these decisions because it’s just looking at the raw data.
And we all know that the AI is looking at the raw data. I can’t give you the whole behind-the-scenes thing, right? Now, there are obviously compliance issues that come with using AI. You have things like sales tax, industry-specific rules, and IRS red flags that often require judgment that AI doesn’t have.
So, again, going back to that nuance of the Starbucks gift card versus just getting yourself a coffee, the IRS is going to classify two of those differently, right? So, if you’re getting a client gift card, that is going to be more of a maybe an office expense, but usually we’d separate that out into client gifts versus a meal. Because the meal is only 50% deductible. Like, if you’re getting a coffee with a networking person, that’s a 50% deduction.
But it’s not necessarily going to understand that nuance or understand it. So, it can have those and create those compliance issues. And then, of course, there’s the overconfidence that the technology is right.
So, a lot of people, a lot of people will just assume that the computer got it right, but the errors are going to compound over time. So, you don’t have someone like having that high-level overview and checking on it. It can really cause issues over time.
One of the biggest things that I’ve noticed is the reconciliation, like the AI reconciliation from QuickBooks. It is so slow. So, you upload your statement.
And I think it took last time, it took me two hours for that statement to fully get analyzed. And then it was just checking everything off. It usually takes me 20 minutes, sometimes even less to reconcile an account, especially if there’s like 10 transactions.
It shouldn’t take the AI like two hours to analyze an entire statement. That’s wild. That is wild business.
So, there’s just a lot of limitations with the technology right now. And eventually, the technology will get better. I’m not going to say that it’s not, but you will still always need a person that can give you the nuance and the context and the confidence in your business.
So, there is a reason that bookkeepers still matter. And that is because we catch patterns that AI can’t. We understand seasonality, client behavior shifts, duplicate charges, subtle fraud risks.
If you’re ordering 17,000 things from Amazon and your AI is just constantly putting it in office supplies, but maybe you’re like, Samantha, I didn’t order this $150 order. AI isn’t going to tell you that. But I might catch it because I’m always asking my clients, hey, what is this Amazon order? Hey, what is this Amazon order? And they might be like, oh, I didn’t order that.
I have no idea how that got there. That’s fraud, right? But AI isn’t necessarily going to catch it. That’s fraud.
They’re just going to categorize it. The bookkeepers obviously will bring judgment. We know when to flag an expense and ask what it’s for or why this is here.
We know when to reclassify revenue or ask clarifying questions. The AI is just going to do the work. It’s not going to be like, oh, hey, I don’t understand what you’re trying to ask me to do.
It’s going to do it, right? Because it’s performative. You put in information and it’s going to perform that task. And then, of course, strategic insight.
The AI is going to give you the data, but it doesn’t explain what it means for growth, your hiring, or your cash flow. And it’s not that deep yet that it can provide all of that really, really good. There’s a balance here.
There’s a balance. And I want to give you that perspective that there is a balance. AI can be a great tool when it’s paired with a bookkeeper, which will help speed up categorization, reduce repetitive work, and help you to spot the anomalies faster.
The danger in all of this is when business owners treat AI as the replacement instead of the assistant. I want you to remember that. The replacement instead of the assistant.
AI is a tool. It is not a replacement. It should never replace anyone.
I could have the AI on Riverside for my podcast edit everything and just get it all done and not even care. But my podcast manager is going to do a better job of cutting out my ums and ahs, of looking at everything and saying, you know what, Samantha, that doesn’t really work here. I’m going to cut that out.
Or if I’m stumbling over my words and I repeat something, of cutting that out. Whereas the AI isn’t always going to catch that. So again, it is not the replacement.
It is just an assistant. There still needs to be some form of oversight. Okay? So obviously AI can process numbers, but it can’t understand the story of business.
And I’ve told you guys this before, that I’m the one who helps you read the story that your numbers are trying to tell you. And that is so important. I want to help you with the story.
That’s why pairing tech with human insight is the safest and smartest path forward. So think of AI like the calculator. It helps, but you wouldn’t want it making your business decisions for you.
You want someone else there to help you make those decisions. Okay? As always, if you found this episode helpful, please, please, please like it, comment, share it with friends, share it on social media, share it everywhere. I love, love, love that you guys are listening to this weekend.
I appreciate each and every one of you who are listening so, so much. If you ever need anything, shoot me a message on Instagram, send me an email, whatever it is. I’m here to chat with you and just talk about things and really just support you and your business.
That’s what’s most important to me. Even if we don’t work together, I want to support you. As always, I wish you the best week ever, and we’ll see you next week.
Farewell fellow travelers.