From Transactions to Transformation: How AI is Changing Bookkeeping and CAS

Welcome to the Accounting for Innovation podcast where we explore cutting edge strategies and insights into the world of accounting and finance. Presented by Decimal and the Radical cba, each episode dives deep into industry trends. Whether you're a seasoned professional or a budding entrepreneur, join us as we unpack key concepts and share practical tips to drive success. Welcome to Accounting for Innovation and I'm excited because today we're going to be talking about AI for bookkeeping and CAs or client accounting services or client advisory services. I don't know what they're calling it these days. But anyways, you know, AI is here and it's helping firms automate transactional work. It's helping to eliminate data entry and scale client accounting services without additional headcount. And I think that's huge for where client accounting services and kind of bookkeeping can go. And in this episode we're really going to focus on the tools that deliver real time categorization, reconciliation and communication. And the funny thing is I've been talking about AI and bookkeeping for over seven years now because I was the original person who was involved in Bot Keeper way back when. So that was like 2018 that that was talking about. And then it was cutting edge. It was really bleeding edge at that time. And all the tools have come so far since that then across the board, whether they be embedded in things like QuickBooks or Zero or their new tools. And so I'm really excited that we're going to talk about that. So with that, I'm going to hand it over to Matt and just ask Matt about his thoughts on AI and bookkeeping in CAS in general, like over the last five years. Because I know like you were one of the early adopters as well.

Yeah. And I think, you know, for us, we, we started decimal in January of 2020. And from the very beginning we have been implementing the bleeding edge, the cutting edge, whatever closest to the tip of that spear in technology. And, and now you're starting to see more and more of that becoming mainstream. And I think this is the best time ever to start a bookkeeping or cast practice. I think the tools make it so that you can not only accomplish the work in really great economical ways, but you can truly start to turn that bookkeeper and that accountant into the advisor. And I think this is a really big turning point for the industry. But the way I think about it, and Jody, this is where I'd like to start our conversation. If you are looking at what the foundation of a client relationship in bookkeeping and CAS is, it starts with the ledger. And the ledger is where the endpoint of all of these operational processes. It's where you pull a lot of the reporting from. But to me, the ledger is that starting point. And you and I both know that. I'm a huge fan of QuickBooks. We use QuickBooks with 6, 700 clients. They are the market winner. They've been doing some amazing things with AI. But I'm also a fan of these new AI native ledgers. And Puzzle is the one that we've had a lot of experience with. But there's also the rillets of the worlds. There's digits, there's Everest, there are a ton out there. How do you kind of look at the ledger space as a starting point for this conversation around CAS and bookkeeping?

Well, I think it goes back to kind of what we touched on last episode about unstructured data, turning unstructured data into structured data. And when you think about the opportunity of what pieces of data you could pull off of anything in a new gl, that's where the opportunity is. Again, you're not locked in to just a description and just an amount. Right. Like, you really have the ability to get all of that data somewhere in a database. And then that means what can you sort through it from what can you, what can you do with that data after the fact? And I think that's where kind of some of the exceptions to the rules come in. Where and when I say that, meaning, you know, right now with QuickBooks, you have to put a tool on top of it to kind of sort through the transactions to pull the exceptions. Right?

Right.

And when you look at some of these AI native ledgers, those tools are already built in the foundation. And I think that's that, that's where like the cool stuff happens, right? It's where, where is the bookkeeping going to come? Where Instead of categorizing 400 transactions, you're only categorizing 10. And now let's put another tool on top of that, that of those 10, it's like double checking those 10 and really popping out only the three or four more, right? So I think it's kind of in the, it's in the do and the review work. That's where it gets that, that's where it gets exciting. Because like, I'll say V1 of AI was like, okay, let's just categorize the last, let's say 20 transactions of the 400. And now when you're looking at like the, the evolution of the tools, now it's like, well, gee, can, like if you have all this other context that it can see, can you get it down to three or four transactions instead of 20? I don't know. But I think that's where it's going, right? And, and then all the reporting that comes right out of it, right, like that you don't have to tag things because the, I mean, yes, they get tagged, but they get tagged. Tagged automagically as opposed to you having, having to put that tag on it so that then when you want to go pull the report, you have to go make sure that everything was tagged correctly. What happens in these AI native ledgers is the tags are automatically like, created. And I, again, I think that's, that's stuff that people don't understand until they've actually been in an AI native ledger. And you're like, wow, I don't have to put in all this extra detail because it just comes with it.

Well, it pulls it in. And that's where when you, when you get. And we can branch out from the ledger too. When you have an AI tool integrated with an AI tool, great example is we've got clients that are on Puzzle and it's integrated with Ramp. Then all of a sudden you have all of this information that Ramp has in paying the bills and categorizing expenses that is pulled into puzzle. And it can pull it into QuickBooks too. But what puzzle does is it provides extra context and information. Whereas in QuickBooks it just pulls it in and it puts it there. Bigger benefit than not having it. And so that's where QuickBooks is really good. But with a puzzle, it can further contextualize the information. And so what it's doing is it's doing a lot of the work that we've all had to do in the past. Well, not me, I'm incapable, but you've done it. But to me, this is where AI is truly helping. The thing that I find is really important here is the understanding that the AI can be 80% accurate, maybe 85, sometimes 90, depending on the client. But it tells you where it's not. And so it gives you the idea of, hey, I know I'm right on these eight out of ten, I need you to look at these two. So instead of you as a bookkeeper or accountant having to look at 10, you, you just have to look at two. And if you think about that, you now can handle five more clients. Because if you have the time and capacity to handle 10 looks, all of a sudden you only need two. Per client. You have five clients. And this is a big change in capacity. And also the thinking through of where people are working and how they can focus on relationships versus the execution and output of work. And that's another thing that AI is doing. And these ledgers and the integration with other AI tools is giving us a lot of that ability.

Well, and I think you have to come back to the fact that we as accountants have to get comfortable with accepting the fact that the computer is getting it right. And I think a lot of times you were so quick to judge the AI and say the AI did it wrong. Right? Well, guess what, Sally, sitting in the cube next to you could have done it wrong as well. Right? And how many times do you do entries get missed or buried somewhere and they're never caught until audit or not even audit. Like two years later, you're just going to look up something and you're like, why is this transaction sitting here? And you can't tell me as an accountant. Well, I mean, I know you're not an accountant, but as an accountant to our listeners out there who haven't found a transaction buried in some g. Buried in some ledger somewhere, and they think, oh, how did that end up here? I reviewed that file and I missed it. And yet they're so quick to like dog on the AI for being wrong when the AI is still probably 95% more accurate than a human who would do that same work. And I think that's a mindset shift that we have to kind of get past.

Well, let me jump in here, because I mentioned earlier, the AI tells you when it's wrong. Do you know what people are really bad at telling you when they're wrong and like that? Even if AI is equally as accurate as the people that work for me, it at least will tell me when it's wrong, which makes it easier. You don't have to hunt for the errors as much, whereas today you'd spend a lot of time hunting on errors, hunting for errors and contextualizing the problems.

Right? So I think that's a mindset shift. And then I think the other mindset shift we have to get through is, okay, if AI is going to do this with all of my clients and I'm going to have four hours back in my day, what am I going to do with that time? And am I capable of being an advisor? And I want to say that because if you're not comfortable doing it, then as much as we tell you, automation is going to help you. Whatever. A lot of people might think, oh, but I don't know how to do that. Or, oh, I can't do that. And I would argue that AI is the tools that's going to help you to learn how to be a better advisor faster now. So if you didn't have time or the partners in your firm don't teach you or the. The pretty cool thing about AI is now there's tools to help us to learn how to be a better advisor faster and to learn how to make a lot of those connections that maybe you had to have years of experience to get to before, but now by using AI, it'll give you the context and you'll be able to learn it faster to get to advisory faster. Because I know from a mindset perspective there are a ton of people out there who are thinking, well, great, if I'm not going to do transactions, then what am I going to do? And. And I don't know how to be an advisor. Right? And I think that's where we have to step back and say, not only is the AI going to help with the transactions, but it's going to help you learn how to be a better advisor faster.

QuickBooks Online and Microsoft 365 are essential tools for modern businesses. But what happens if your data is lost or compromised? Syscloud, a proud sponsor of the Accounting for Innovation podcast, provides advanced backup and recovery services as well as ransomware protection and anomaly detection, ensuring your business can continue operating even in the face of disaster. Syscloud peace of mind for your business's critical data. Visit syscloud.com to learn more. That's syscloud.com and now back to the show.

So I'm interested, Jodi, if, if somebody listening to this is thinking, gosh, I'd like to be an advisor, I'm not sure how to do it tactically. Walk them through steps they can take to feel comfortable, maybe to learn and upskill themselves to becoming that advisor within.

A cast Practice Start thinking about using either ChatGPT or one of the other large models as a strategic thinking partner. Start thinking about it as your business best friend and start asking it those kinds of questions. And, and you know, depending on what version you have and I would say if you have a paid version, you're okay with uploading some and it depends on where you are in your firm too, because I know there are some firms who still like, they're not allowing it at all but you know, ask it questions about a P and L statement, right? Like you can give It a comparative P and L statement and you can start just asking questions about it. And if you don't know the question to ask about it, the best question to ask is to say, if I were a CPA advisor, what should I be looking at on this P and L and balance sheet? And what should I bring that to my client? What should I bring that to my client? Right. And so I think what's cool about it is if you were always afraid to ask those questions to the partner or the person in your firm, you don't have to feel stupid with AI, you can just ask it all these random questions and like you, you don't have to worry about not knowing.

Right. I, I think that that is that kind of curiosity. Accountants can be really good at being curious. Being curious at using these new tools, at testing their capabilities, I think is ultimately where people are going to become more comfortable, learn, and then start using them on a more regular basis.

Well, right. And the thing is, is like, if you don't know where to start, I love asking AI where to start too. Like, you could always just say, hey, like I want to upskill myself and become a better accounting advisor. Where do you suggest I start? And like, it'll, it'll give you some.

Yeah, just listen.

So, so I think that's, that, that's the thing is like a lot of people are, haven't really gone into play because they're afraid they're going to break something, something or they don't know. And that's where you just start learning. And that's why I really like saying talk to AI like it's your business best friend. Don't think you have to learn how to prompt. Just start talking to it and see what it gives you. And then it'll help you allow it to become a better strategic partner of yours. And then you'll learn more advisory skill sets. AI is going to do most of the transaction work. I mean like it truly is will eventually. So how do you stay relevant? And that's by really, you know, changing your mindset and starting to learn the other tools to keep you there.

I agree. And, and I think too it's, it's relooking at the technology that you put in place for your clients. And as you do that, putting in the tools that also provide the ability to use AI better. You know, the integrations, the tech stack that you put in place either makes that work more capable for AI or harder, just like it does easier, harder for people. AI is basically the same thing. If it has no access to data, it can't do anything with data. And so looking at putting together a tech stack and what we use at decimal, our preferred tech stack is quickbooks or puzzle for a ledger. Probably had RIT later this summer. Then bill pay is ramp or bill.com expenses are ramp. And Brex payroll is gusto and rippling. These are the technologies that we've really found integrate well and provide a good foundation for then layering AI on top. Or they're already using AI in a way that is making the work better.

Yeah, and, and I think the other piece of it is, is like, I think there are still people who are overwhelmed with tech stack in general. Right. Like they don't know where to start, they don't know how to begin. And I think the idea is start just. Well, first of all, just start. But then the idea is, is like you don't have to know every, every program. You just have to start to get to know a few and then build from there. Right. So like, you know, you just rattled off I think 10 different softwares or something. And if you're just starting, that can be a little bit overwhelming. But the reality is, is, you know, start with your, start with like obviously your GL and then, you know, add one or two and then keep going from there. Because I think if you've never really had a CAS practice, you know, the tech stack can be a bit overwhelming.

Well, and what I would say is start with the gl. But when you find your pain point, maybe you've got a bunch of clients and bill pay is painful. Maybe you have a bunch where payroll, maybe you have a bunch where expenses, whatever it is, whatever problem set you run into, then go find the right tools and then the next time you have another one, go find that tools. You don't have to dive in as deeply as we have or use all of those. And I'm certainly willing to always talk about tech stack and how we use it, but I would say in general, follow the pain. Wherever you feel pain or you feel like you're doing a lot of work, see if that's where you can start to layer in better tooling. That makes your life easier.

Yeah, I mean, I think it's really important because I think a lot of times people get overwhelmed with the tech piece. And I think the good part is when you think about AI as a whole, the cloud, we really had to know the tech. Yes, it was. Well, I'll say it was easy. It was easy for me. I know not everyone Thinks it was easy, but you kind of learned how to connect all the digital plumbing. Right. I think with AI, it's a lot more user friendly. Right. Because it's more like an iPhone where more approachable. Yeah. It just kind of works versus, you know, the cloud, where a lot of it was, oh, you have to learn it. Right. And so I think, though I would argue you still have to get your hands dirty and you have to get in the tools. Right. Like, it's not like you can just do a webinar and say, oh, I learned it, because I think there are too many people who think they're going to learn AI through osmosis. You still have to touch the tool.

So you only get. You get the best experience by doing, not by just reading and learning. And I think. I think as we wrap up this episode, that's a great way to do it, is saying, hey, and I think you're probably going to hear this theme throughout our season of AI, and my guess is, Jodi, as you and I get into season 25, it is going to be the get your hands dirty experience is the best teacher. These are new technologies, new things, new tools, and be willing to try them, be willing to test them. Set aside time every week to just playing around with stuff. That's a big thing. Develop the habits that give you the room to do it. And I think that's how we build the tools. And with that, I think this has been a great episode. It's a good starting point for people with cast and bookkeeping on the mind to think about how do I build the practice, how do I make it easier, and where do I go from here?

Awesome.