392 Bookkeeping Tech Predictions for 2025!

Today, team, we're talking bookkeeping stuff, accounting stuff, the ledger. What should we expect from our tech partners in 2025 from the broader trends happening around AI and general tools? Anything that's going to make that firm, I don't know, more profitable than it was in 2024, like tech that we need to be planning around today that might be just around the corner. Today we're going to dig into the biggest things that happened in the last 12 months, like the trends that everybody needs to be on board headed into 25 and, and get the old crystal ball out to consider what may be just around the corner. Come on in, pull up a chair. Let's talk bookkeeping tech trends for 2025, Gang. It's a snow day here at at Unfirms hq. Just got in from absolutely freezing my feet off because for some reason I don't have snow boots. Just one of those things you don't use enough and then you need them and you're like, I know I have them, but I don't know where they are. So we're going to work that frostbite off today by talking bookkeeping tech. Okay. Yesterday we did the same for tax tech. Recapped the biggest things of this last year and what's to come for next year. I'll be totally honest, I just rip the band aid off right now. While there are interesting things happening at accounting and bookkeeping tech, I would say on a scale of 1 to 10, that 2024, the hype level was maybe an 8, maybe even a 9. Just with all the tech stuff that's happening in general. But what would you score 2024 on the Spectrum of like tech developments for bookkeeping firms? I'll be honest, I'm going to give it like a four. And a four is a bummer as it is. But a four when it should have been at eight or nine, that's just a double bummer. You know what I mean? In fact, I would argue if you take Keeper out of the picture, which I think is the, I mean is the biggest category shifting bit of tech we've had for bookkeeping firms in the past few years. And one of those tools that just has like an absolute ridiculously fervent level of support right now. Let's link Keeper down in the show notes. If you take them out of the picture, boy, I don't know the really much at all changed around the ledger, but let's get straight into it. What are the most interesting things that did happen in 2024, good or bad. And not unlike the episode we did yesterday on our tax workflows for US Tax, the core software that people use to produce tax returns is unchanged for the last 25 years. And so when we talk tax tech, we're actually talking about all the things that happen around that tax prep software. Interestingly enough, for bookkeeping tech, for accounting firms that are just doing a volume of month end closes for their clients, the story's kind of the same. We're actually not talking all that much about the interesting things that the ledger platforms themselves are doing, which is kind of a bummer. We're talking about all the interesting things around it in the ecosystem. If the ledger systems were doing interesting things, I mean, for sure we would be shouting that from the mountaintops, right? That'd be really exciting. But actually not unlike US tax tools these days, our ledger tools, they're pretty mature. And honestly the only ones in the US really with any amount of meaningful market share in my opinion are QBO and 0. And I mean 0 is just off a cliff relative to QBO. But then even after Xero, if you're talking freshbooks or anything like that, is off a whole other cliff to where I think honestly QBO and Xero are kind of the only super relevant to small accounting firm ledger platforms out there. You go up market more. You're talking like Netsuite intact, that sort of thing. But what has QuickBooks and Xero shipped this year that's fundamentally changed your workflows? Insert, insert crickets, not MP3. Kind of too bad, right? So most of the stuff we're talking about is around the ledger. Although we will get to crystal ball out and explore what are some interesting things ledger players ought to do that I would love to see. All right, number one, and this is actually a big thing that's kind of sort of related to bookkeeping firms. Accounts payable, accounts receivable and purchase order processing. These three things are the most disrupted by AI things that have happened in the entire accounting industry in the last 12 months. It's been a long time. I mean people have been saying for a long time with OCR technology, oh, we finally don't need humans to process documents anymore. And if you do any level of this yourself, you know that has just fundamentally not been the case. Particularly in, in really high stakes applications where like ar, ap, Purchase ordering too. Like that is just a fundamentally high stakes thing where if a piece of software reads an extra digit where there shouldn't be one, or misses the period between the dollars and the cents all of a sudden you're paying a bill, you know, $40,000 more than you're supposed to. And the notion that anyone has ever just completely covered their eyes and put the invoice in, let the machine go, brrr. And then let the same machine then pay that invoice, like it's just been ridiculous. Like that's absolutely not been the case. The hard thing too is you have the age old problem in accounting firms where you pull in really good OCR technology that does, let's say it gets 99%, although it never does. But that's usually what you see on the, on the box it gets 99%. But does it remove your need to check it? No, it doesn't. And there's a scenario where maybe it removes the prep process and so you're just doing review. And I think that's how it's framed. But the reality is attack shops know this all too well. The reality is as long as there's a prep process still the preparer is going to do the same thing that they always did because ultimately they're accountable for it going to the next stage correctly. So they're still going to check the work. So unless that OCR process process is like somehow removing an entire step fundamentally where the human being doesn't even touch it until it gets to the review stage, that hasn't really saved any time. Now what's changed in the last 12 months? Well, the really big unlock here is two things. One, it's AI Vision models and their ability to fundamentally visually process things like in a like vastly more accurate and more intelligent way than OCR ever could. And we are a huge step beyond already what OCR could ever do. And it's gotten pretty cost effective pretty fast. And what the best platforms are doing are not just relying on that better vision so it can pull stuff more accurately. Even pulling handwritten text for the first time, which was not possible with OCR to any degree of accuracy. They're not only using these vision models, they're also now using large language models to layer in a bit of reasoning to say does this make sense? Which is actually a really, really big deal. Because if you think back to what OCR has done on document extraction, oftentimes what it will get wrong is like it's one letter off on a vendor name, so it creates a new vendor. No human being would ever do that, right? Like that's such an obvious connection to make that, that's one letter off pulling in an invoice where the invoice date or the due date or something like that is a year off or it's 20 times the amount any past due to invoices have been like there's certain things here where no human going through that process would ever be like yep, nope, just send it through. Like it's just such a plainly obvious thing. This is a really cool application of large language models these days where we keep, you know, putting on LLMs, this incredibly high bar of like they need to match human levels of intelligence to be of any use. It's like, no they really don't. Like if you think about plugging them into applications like this where there's ways that you can hard code things to look for but if you have an LLM that can just in a more open ended way be like, is this reasonable or not? Those LLMs don't have to be extremely intelligent. Some level of intelligence is really, really useful and that's actually what's making these systems even better than just doing better amount extraction. So of all of the things that happen under the quote unquote accountant profession or whatever you label the really vast set of tasks that go to accountants in industry and in public and all the different hats that accounts wear, this is the area that has been most disrupted and there are genuinely people actively being displaced by this right now. It's not to say that these accountants cannot then go do other work because at the end of the day if an accountant's job is to sit there and just process ap, just process ar. I think people probably would have said for a number of years now you're living on borrowed time. That's not a particularly like big brain thing. It seems very suited for computers to be able to help out with this stuff. What has surprised me is the complexity of the scenarios and the use cases where these tools have come in and are frankly really, really good and can even make these associations that require kind of like several steps of logic and are not just a compare this to that perfectly map the item into the system like we would think with OCR platforms in the past and had to be very kind of specific and on the rails. I've been incredibly impressed with some of the things that I've seen around. Purchase order processing is one that's sometimes a little bit more open ended. I've been impressed at just how intelligent these systems feel and they still suffer from a bit of the same problem where it's like, well if they're 99%, I still have to check in the case of the 1% but I think at this point they have fundamentally eliminated the need for anything but review, which I wouldn't have said was the case in the past. Like it's, it is on an entirely other level that is I think, the biggest thing that has changed. And in 2024, if there is an, an opportunity or an insight here for accounting firms, it is that I do think that there's actually a bit of an arbitrage opportunity here where you can come in and like really turn the process of a small business on its head. Because SMBs are not super fast to lean into this stuff. New tech, weirdly enough. Like this sort of thing usually gets plugged in at the enterprise level, at the mid sized company level first, and eventually trickles down to SMB. But so often accountants are the ones injecting new tech into SMBs and introducing them into how, like, like how to build a proper finance tech stack, that sort of thing. There's an opportunity here. I had clients who genuinely had people like FTEs worth of people who were processing invoices and doing receiving and stuff that was very important because you had inventory management and all that. And with the tech we have today, we could have come in and wiped out effectively, you know, 80 to 90% of these people's time. And in every case I can think of, I mean, I think there's other things that the business owner happily would have had these people do. There's a real almost obligation or social pressure to acknowledge the displacement of people in conversations like this. Almost to like sugarcoat this, to say like, don't worry, don't worry, they won't be replaced. I think it's kind of wasted effort and I think like we need to acknowledge it for what it is. It is us trying to make us feel good about something. The reality is people are being displaced by AI every day. Not as much in our space and the stuff that we do yet, but definitely in anything audio, video. Like there's, I mean there's so many examples of where AI is coming in and doing things where people are like, well dang, that was my job. And while that may not immediately impact you yet, oftentimes this impacts our clients, which is just as significant to us. Like if I had specialized in working with voice actors, oh my gosh, is it a rough time for voice actors right now? Because AI voice stuff has become so, so good. Even if you are just tangentially related to an industry, maybe you work with the agents for those people or something that's, you know, connected by some second or third degree connection. There are ripple effects to that stuff that affect all of us. And so it's like I open 2025, we're not as dismissive of AI as it feels like a lot of us still are because stuff's changing. And the biggest place where that impacted accounting was ar, AP and purchase order stuff. All right, second biggest, biggest. I don't know if trend's the right word, but thing that happened in 2024, Intuit finally stuck a fork in desktop. Like in a, in a way that feels like it is for good. The writing was on the wall for years, but Intuit took another really big step in their seriousness of shuttering desktops, QuickBooks desktops capabilities with the exception of QuickBooks Enterprise. And Enterprise feels inevitable to a degree that they do the same. It is something that we knew was going to happen, but for it to actually happen feels significant. And a lot of, a lot of companies are still very desktop reliant. A lot of them are using desktop for you know, volume payroll processing. And you know, it is as if, I mean imagine if somehow overnight all the accounting firms had to go paperless. And it's very easy to say like why is anybody still doing stuff on paper? But it's like anytime like somebody snaps their fingers and says it has to be done a different way now, like that's a big deal. Like there's still a lot of people that are using that platform and it is kind of forcing modernization in some ways. Number three, I would say there's been a surprising lack of new sort of from scratch AI driven solutions. Whereas on the tech, on the tax side you actually see quite a few new apps that are AI first companies that are doing new things thing doing tax things in novel ways courtesy of AI technology. You see AI research platforms, you say see AI tax intake platforms. But on the accounting side, man, I mean I don't know if I can put my finger on a single company that is like an AI first company that has brought a new novel thing to the market. There have been abundance of companies who have raised money and are saying like oh, we're automating this or that aspect of accounting, but it's literally just a slide deck. And it's literally like extracting things from documents, a very table sticks stuff. I'm only interested in that thing if you can put it in the context of actually completing a monthly close and saving somebody a pile of time. And for all of the early stage companies that are out there, like I still couldn't put my finger on a single AI first company that is doing something meaningfully different in these workflows. The exception being the first one that we talked about where ar AP platforms like, have now taken a really big step. So of them, a few of them, most of the legacy ones, honestly, like I. I haven't seen that much from. But I think if we had, if we could go back 12 months to the beginning of 2024, I think we would have probably hoped that more really novel AI stuff for the accounting ledger would have come around. But it certainly hasn't. This episode is sponsored in part by Live Flow. It is. It's the time of the year where we make resolutions for what we wanna do better in 2025. And one of those resolutions might just be upgrading those client touch points, improving the client experience. Because at the end of the day, everything kind of boils down to that and how premium your clients perceive your service to be. And one of the most automated ways to create really valuable touch points that are like actually connected back to the accounting ledger, but it's like live. It's Live Flow, guys, Gang gals, squad, Live Flow. Originally the secret sauce was they were a connector for your spreadsheets. They would connect QBO and Zero files to your Google sheets or your Excel. And they still do that very well. But they now also have like an online dashboard or through all those same connections you can create like super quick and easy, not overwhelming. If you don't want them to be dashboards, it actually feed in live data from the accounting ledger, either on a continuous basis or just a snapshot. So I could like send a roundup to my client over text every single morning. It opens up all sorts of possibilities. But the tool itself isn't a bunch of work. That's what I never liked about reporting systems, is you log into that thing and it's just a lot. Customizability is a big thing here. LifeFlow always has really nice templates you can swipe, but you can also build it into whatever you want it to be. Because ultimately how clients value our service, it really boils down to those touch points, what we put in front of them and what they get out of it. To learn more about LifeFlow, check out the link down in the show notes. This episode is sponsored in part by TeamUp, the people who help you find talented accountants in the Philippines but hire them directly. Very common progression I see from firms who lean into offshore hiring for the first time is they turn to like some offshore megacorp and it's like this big skyscraper where they put folks in there and they're maybe scared about security things and all that. Understandably, the first time they work with people offshore and they're like, oh yeah, give me this big sanitized environment where they like strip search everybody before they go into work at the beginning of the day and they leave. And it's like, this is fundamentally more secure than the way that we do our work onshore. Like, I can understand how this works, but before you know it, you're paying this group, like you're paying them as much money to employ those people as you're actually paying the people. And when 50 cents of every dollar that you pay goes to some other company instead of the actual person you employ, that's just a bummer, man. That's just a bummer. So a lot of firms I know transition to just hiring people directly. That is what Team up specializes in. No, no middle people. You're just engaging that person to work with you the same way you would somebody onshore. And 100% of that money is going straight to the person. One hundo dollar out of your pocket, dollar into their pocket. Guess it wouldn't be a dollar. It'd be whatever the local currency is. You get the idea. To learn more about Team up, check out the link down the show notes. This episode is sponsored in part by Maker's Hub. If you're thinking about getting into some new, more profitable services in 2025, one that I cannot recommend enough is advising your clients on their kind of short term cash flow outlook. Oftentimes it's a prerequisite to be able to do more long term planning, which we want to jump into. And then we realize, oh, they're not going to make payroll on Tuesday. So that's kind of what they're focused on. That was the story of like my evolution. We tried to do long term planning, that didn't work. Then we started doing short term planning and then we're like, I kind of got to have an under better understanding of your AP setup. So then we started doing bill pay and that kind of created the loop and opened up a huge advisory opportunity. And AP right now, funny enough, is being like completely reinvented by AI. It truly is because we can actually do much, much more automated extraction than we've ever been able to do before. Makers Hub, they're pushing it farther than anybody else that I've seen today. We did a demo day on them a while back about how AI is kind of revamping ap. I think it's now the most watched accounts payable related video on YouTube anywhere. Because it's like really compelling. It's really impressive if it's something you're looking at for 2025, take a look at Makers Hub. They recently rolled out a bunch of stuff around purchase orders too. Really impressive tech link down in the show notes number four, Intuit made everybody mad again. And then we kept using into IT products. Uh, let's just, let's just keep going. Number five, we got AI assistance for the accounting ledger, AI assistance for the accounting ledger, like chat GPG for the accounting ledger and nobody cared. So first to bring it to market was digits. Next was QuickBooks with Intuit Assist and then was Zero with Jax. Probably the order we would have expected. A third party kind of big brain AI company does it first and then Intuit does it and then Xero does it. And I don't think it's changed anything for anybody besides a few people that run a car wash and they're like, I'd rather work with it via this assistant than the actual interface. I can see the case for this being useful long term as more computer stuff just in general goes conversational instead of using the user interface. And for a lot of what we do, that seems hard to get our heads around the notion that you're just going to have a conversation with something that's going to do that work instead of actually use the interface. The best analogy I can make here, which kind of connected something for me was switching to conversational is not a replacement necessarily for the underlying need for that user interface, but it's something that we do every day when we work with other humans. So if I run a team and if I talk with those people and I give them instruction, I say go do XYZ and then report back to me with this. That's actually an example of something that's not too different than having a conversation with a piece of software in order to get it to go and do a thing. As many problems as there are with anthropomorphizing AI and sort of humanizing it, the reality is the way that we work and the way that our brains are wired is to work with other human beings. And so sometimes that's a helpful construct to be able to frame something that an AI can do for us in the terms of, well, what would a human do? Because we're used to that, right? So I can see a future where maybe once these, these assistants are much, much more powerful. And, and an important turning point there is are more Capable of doing things rather than just talking about things. When we're getting work done, we're doing things. I hire people to my team because they can get the work done. So an assistant that can talk about the underlying stuff, not particularly helpful for me until it gets to the level where I can tell it to go and do XYZ and some of those, like Jax0's assistant, for example, you could have it do stuff like send in voice reminders, that sort of thing. And some of that is cool. The reality is it just has not become normalized yet to work with software in a conversational way. And I do think that's changing. Sam Altman OpenAI CEO talks about this is something I've always kind of felt like, even like you will have a day where a super groundbreaking technology drops. And arguably, you know, some people would tell you that we actually got artificial general intelligence a week ago with OpenAI's announcement of O3. And there are some well known benchmarks where they say basically once something tests beyond this threshold, we can consider this AI to be AGI or artificial General intelligence. A level of intelligence where you can set that AI loose to work on that thing kind of on its own and it performed better than the average human is kind of the idea of artificial general intelligence. Well, they did some testing on this new model and it performed beyond human levels. Now the devil's in the details there. On that benchmark, they basically spent as far as we could figure, about 1.6 million on compute in order to pass that benchmark. So for the low, low price of $1.6 million, you can have something that is slightly better than the average human. AI is getting cheaper at a faster rate than humans are getting cheaper. And so that's, I mean, the idea is ultimately that becomes much more cost effective. But you know, ChatGPT's advanced voice, the ability to have an incredibly human conversation with something and it feel like you're talking to an actual human being. We've had this for a while now, but most people have not experienced the power of it and just how compelling it is, like it's unbelievably good. ChatGPT, what are some ways that you could see a voice assistant being integrated into an accounting ledger in the near future? Integrating a voice assistant with an accounting ledger could streamline tasks like recording transactions, retrieving financial data and generating reports. You could easily add expenses or check balances using voice commands, making it more efficient. It could also provide reminders for upcoming payments or deadlines, and even answer Financial queries on the go. Okay, that feels more or less like what some AI assistants already do with accounting ledgers. But how could voice make this easier for me? And if you can just, just talk to me like, like an old Italian man. I don't know, maybe just slightly grumpy. Sure thing. Imagine this. You're sitting there talking to your ledger like it's your nana. And you just tell it, hey, record this expense and boom, it's done. No fiddling with keyboards. You want to know how much you spent on tomatoes last month, just ask. It's like having a chat with a friend, but the friend keeps perfect track of your finances. Easy peasy. Okay, what, maybe as a Bostonian now tell me, like, what would the actual inputs for this be? Like, how would it access the files, where it would like the stuff that it would need to be able to actually make those determinations? Like, would it be fetching that from my computer? Like, how would that actually work? Alright, so picture this. You got this voice assistant, right? It's connected to your accounting software on your computer or even in the cloud. You say, hey, love this expense. And it pulls up your files, your spreadsheets, whatever you got. It's all integrated, see. So it fetches the data, makes sense of it and updates everything in real time. Makes life wicked easy kid. You get the idea. You have these pretty novel, powerful AI tools now and ChatGPT has gotten really good. Advanced voice, I think is incredibly novel, but people don't even know about them yet. Over the holidays, spending time with family, like, nobody's talking about this stuff. We just got a version of a ChatGPT integration with Siri, the iOS smart assistant. And I think that makes it a little more mainstream when it's like, my mom can be having a conversation with it and she's like, oh, I heard about a chat, a chatgpt. And you know, I've got, you know, some non tech people in my life using it now for the first time and they actually keep coming back to it because they're like, wow, this is actually really helpful. And I can send it pictures and on the app for ChatGPT now, you can even stream video as you're having a conversation with it, so it can see what your phone can see and it can like walk you through, you know, fixing something in the washing machine or something like that. That is incredibly powerful, novel technology. But even though we have it, we're still a long ways from it, like becoming normalized and used every single day. And so on the subject of like these AI assistants working with accounting ledgers. And it's probably going to be a long time before people are like, I would rather talk with a thing to get that thing done than log into the interface itself and do it. Particularly for power users, which are very different than your normal small business user, you and I already know how to do that thing in the interface. And so it's in many ways more like habitual and like natural for us to just hop into the interface and do that thing as opposed to asking an AI to go and do that thing for us. But I do think that inevitably changes. And the example I always come back to is in Microsoft or in Windows, if I want to enable night mode right now, you can just hit the little shortcut and tell Microsoft Copilot or Windows Copilot enable dark mode or night mode or whatever it's called. And it just does it. But you know to like go into the settings, you don't know exactly where that is in the settings. And you may search in the settings and sometimes it finds it, sometimes it doesn't. But just fumbling through those settings is a real pain. But telling the AI assistant to do it, that's just fundamentally better. And once you do that once, you will never go back to the settings to do that same thing again. So there's going to come a day where we learn to trust these assistants and we rely on them and they become the easier way to do that thing rather than something that's buried under a bunch of menus. But we're not there yet. And even though we have some AI assistance for the ledger, which is awesome, and I applaud the companies who have put this out there because you kind of don't know is this going to be groundbreaking or not until you've actually got it in your hands. We have it now. It isn't groundbreaking yet. But as conversational AI becomes more the norm, I suspect the notion of working with these things over voice will also become more of the norm. Number six, last one, kind of the biggest trends of 2024, arguably the ledger platforms Puzzle and Campfire, we can link each of these in the show Notes became somewhat legitimate ledger solutions, and I say somewhat legitimate, not like to demean them because they're both on their own in a vacuum. Perfectly fine ledger systems that you can run your entire business on. The problem is the ecosystem advantage that QBO and Xero have. So if I'm a firm that's managing a bunch of month end closes and keeper, those platforms are not an option for me because they don't work with all the other tools that I use to run my bookkeeping practice. And so it's like the platforms themselves I think are perfectly fine, but because they're not supported by all the other tools that I use, they aren't really an option for me. And this has classically been the real struggle of overcoming say QuickBooks advantage in the US is just the vast ecosystem of tools that connect to QuickBooks. And it's not going to be until somebody finds a fundamentally superior way to manage the ledger that has to overcome that ecosystem advantage. Like it's going to have to be so, so, so good. So much better for somebody to pick that tool up even though there is no ecosystem advantage with it. Right. Puzzle and campfire to me, like they don't rise to that level yet. Hopefully they do someday because QBO and 0 need that competition. But the tools themselves, like the ledger alternatives themselves in a vacuum I think are better than they've ever been now, but for most people still don't rise to the level of being enough better to overcome the benefits of the whole ecosystem that connects with QuickBooks Online. Okay, enough talking about the past. Let's talk about what's to come. 2025 Jason. That's what I'm here for anyways. Get down to it. Okay. Number one, going to see more tools doing deep ledger integrations with QBO or Xero. And Keeper's kind of led the charge here. The novelty of Keeper being that they sync in like all ledger transactions and let you manage a month end close from inside of Keeper rather than having to log into a whole bunch of different account files like client files. We've seen a version of this now from say Client Hub Financial Sense, like where they're syncing some transactions and doing like kind of an uncategorized expense management sort of thing inside the practice management system platform. But the reason why Keeper is the leading month end close automation platform now, which is very different than being a practice management system, which is what Client Hub and financialsense are. But the reason why Keeper is caught on for automating month end closes is that it literally syncs the entire general ledger, letting you build out kind of a task a to do system around the transactions themselves. And when it has all that information it can do other stuff like automating your zoom expense. In the past it's always been classified this way, but now it got classified this different way this month. Do you want to change it also? Just other smart things like you Know how in every ledger, if you're looking at a balance sheet or a P and L, then you drill into the account, it's just a list of those transactions by a flat list of those transactions. No hierarchy by date. In Keeper you drill into say an expense account and the next hierarchy is all the transactions grouped by vendor. So I can see the total spent on Zoom month over month before I drill into Zoom and see the individual transactions, which is just a fundamentally better way of viewing transactions in my opinion. Because what's the first thing you do when office expenses 3x this month you drill into it and you compare it to last month and you're looking for any big outlier transactions and you may see like you know, four purchases from Costco one month, one purchase from Costco another month. Or say you're comparing software spend month over month, maybe there's a better example. And you're like, why is it twice as much as it's been in the past? Once I drill into software expense or wherever you bury your software, the next level that's most interesting to me is the payee, who am I paying the most to like the total. And that could be six transactions, could be one transaction. Keeper does that for you. That's super nice, a great, nice to have. But the general ledger is kind of, as we talked about, it's kind of what it's going to be now. Like there's not a ton of innovation happening at the general ledger level. And so all the workflow automation is happening around the actual general ledger tool itself. It's happening when it comes to, you know, how you review what's in there on an automated basis, how you request stuff from clients on an automated basis and then push what they submit back to the ledger with the classification. It's stuff like that. It's not inside of QuickBooks, it's not inside of Xero. It's. It's all the tools happening around that. And the more data that the tools around it have, I. E. The more transactions they sink in, that sort of thing, the more, more interesting things they can do with that data. And a lot of it is frankly, I think people like following on what Keeper's doing. Although I will, I will say one other thing there, take practice management systems for example. Let's say a practice management system is trying to build a super compelling email assistant where the replies suggests to emails with like an AI assistant where a client sends you an email and the AI is going to draft a reply to that email for you. The quality of Those responses come down to the access like how much context it has. Oftentimes your email client like if you using Outlook, it's not connected to anything intelligent. Right? And this is why email inside of your practice management system has a much higher ceiling for how intelligent those AI driven responses could, could be. Well, most practice management system systems are still not connecting to accounting ledgers and to me this is an absolute no brainer. You want that information to just authenticate a client to their QV ledger, to their Xero ledger so that when the banker asks can you give me a copy of the profit and loss for last month or, or can you tell me why office expense was more this month than last month for it to automatically be able to go out to the ledger and pull the detail and draft a response or even just present to you the detail that makes up the difference. Man, that's invaluable if that's already there waiting for me when I hop in to reply to that email. And there's a whole bunch of examples of scenarios like that where I think it makes sense for more other tools to connect to the accounting ledger, maybe sync some of that information into their platform so that they're able to do more intelligent things with it. Month End Close Automation is the most obvious example there. If you're using an app like keeper to automate 100 kind of client month end closes from a single place. But even for practice management systems, I think it makes sense to pull more of that ledger detail into their platform so that we have less of these siloed places where we're doing work that can't see into each other. Right. Okay, second thought, I wants to come for 2025. I think practice management systems will begin to specialize down the path of. We talked about this yesterday. Are they great for tax firms or are they great for bookkeeping firms? Are they trying to do like half big version of both? So if you look at what Keeper does it is Month End Close Automation. We got a couple practice management systems that lean a little bit that way, like Financial Sense and Client Hover, the ones that come to mind more than any of the others. But the third item that I have here of what's to come in 2025, I think keeper becomes a practice management system. That seems inevitable. They actually just shipped some features for tracking your tax projects inside of Keeper now, which basically means if you the direction they're headed. I don't, I'm always hesitant of a version one actually being good enough to use. But you see the Direction that they're headed is they have a bunch of firms that do both tax and bookkeeping, but they can only use Keeper for the bookkeeping side of the business. Well, if they could run all their tax projects through there too, wouldn't that be nice? And then maybe you don't need the other practice management system as a much bigger thing to bite off because then you're talking, you know, time and billing and a lot of other things your practice management system does that Keeper doesn't do yet. But Keeper, who's kind of the, the elephant in the room for, for bookkeeping firm's tech stack because they are arguably solving month end close automation in a more compelling way than anybody else. The practice management systems all see that. And so there's a version of, well, could we do like automated quality assurance in our practice management system and say that we're now the best practice management system for bookkeeping firms? Right. You don't need those other tools because we're now going to do a version of that in our platform. Like we talked about yesterday. You're going to see practice management systems getting into tax intake and getting more involved in the tax workflow. But for a, I mean it's enough of a, enough of a hard thing to imagine that a practice management system will nail tax workflow as it is to think that a single practice management system will nail both bookkeeping workflow and tax workflow. Like do this deep ledger integration and do a really good version of tax intake. Is anybody going to do all those things really, really well? That seems exceedingly unlikely. So the net result I think is we have practice management systems that are more specialized to a specific type of firm, which I like, I think is a good thing for the accountants. I don't know if it's a good thing for the practice management systems. It's a good thing for the accountants because right now there's a ton of practice management systems out there and new ones actually coming up every year. But what I want is just the short list of like, okay, here's what I do. I'm this size, I do bookkeeping and tax, or just do bookkeeping. Like what are the three best options for me? And the way we're trending, where they become more specialized, those three options are now even better for you than they've ever been before because they're built with a very specific firm in mind. I think that has to be the future of the PM category because these workflow tools like Keeper, like the tax intake apps that are getting really good. They're in some ways they're starting to take the PM's lunch a little bit. Because when those things start shipping PM features like if Keeper becomes a pm, the PM tools have to respond to say, well, how are we going to deliver some sort of month on close automation functionality? Because that is saving firms a bunch of time. Like that new level of sort of workflow tool that's doing deeper work than a practice management system does. This episode is sponsored in part by Cloud Accountant staffing. Hey, scenario. What if you had this buddy that was running another accounting firm and they stumbled into this thing and they like, they stumbled start hiring these really good offshore accountants in the Philippines and you're talking with them one day and your buddy's like, hey, I got some really good people here. You want me to hook you up? You want me to find you a few accountants? What would you say? Would you be like, nah, I'm good. I could find accountants all day long. I don't need more accountants in my firm. Poppycock. You'd be like, tell me more. That's what working with cloud accountant staffing is. Because the team behind cloud account staffing, they were literally just an accounting firm that started hiring people in the Philippines, had really good luck, like I did, with the quality of who they picked up and then they just started doing it for other firms. It's not, it's. It's kind of as simple as that. If you're looking for somebody who understands you just works with accounting firms, not some rando sending you cold DMs on LinkedIn, a group that actually does this stuff just with accountants. Check out Cloud Accountant staffing link down in the show notes. This episode is sponsored in part by Carbon, the practice management system. Funny story about Carbon. So we did our, our 2025 practice management system recommendations. There's like 40 some practice management systems out there. And we pared it down to a list of, was it 7? Some much shorter list where I'm like, these are the only practice management systems that I recommend. And doing that was a big undertaking. Cause we looked at basically six different sizes of firms. And for each of those different sizes, firms have different service mixes. So if you did tax, only if you did book bookkeeping, only if you did tax and bookkeeping combined. And for all those different scenarios, we gave a top recommendation and two secondary recommendations for the practice management systems that are best suited for those types of firms. Now only two platforms got five top recommendations. One of them was Keeper, which honestly is not even a practice Management system. It's like month end close automation. The second was carbon. Carbon's really good man. It's like it's one of the only tools out there that really raises to the level of if you are considering tech change in your firm, honestly, practice management system or otherwise, you have to see what carbon doesn't understand what it does because if you go down the list of really, really smart like thought leader accounting firms right now, Chad Davis, great example comes on the pot a lot, a lot too much. A lot of really smart firms are running on carbon. And sometimes honestly like that's what I need. That's all I need to see. Do I really need to like outbid big brain these other firms that I look up to or do I just say okay, they've done a whole bunch of due diligence. I still gotta do a bit, understand what works for me. But there's, there's something to that. Learn more about Carbon, Check out the link down in the show Notes. This episode is sponsored in part by corvy tax firms. Please do not go headlong into tax planning season without the right tool for the job. You are flying blind, you are wasting time and you're probably underselling the value of, of your work. Core V to help you find like a bunch of tax saving strategies for clients. Like not just maximizing retirement contributions, but a whole bunch of opportunities. But the problem for most firms is we just, we undersell the value of our tax advisory work across the board. You look at small tax firms, the ones that are absolutely crushing it. People we've talked to on this podcast that are working like half time making several hundred grand a year, they are making that on the back of charging top dollar for tax advisory. And I honestly didn't know how to do it until I saw a really good tool. Prove the value of your work and we kind of forget about it. Like we understand like yes, the stuff that we do is very valuable. But we forget about just how much money we can put put like we regularly put back in our clients pockets through tax savings and tax deferrals. And then we go out and you like bill the time. You know it took you an hour and a half to do that. Here's this $400 invoice for that 40k in tax that we just deferred. That mindset shift of how to better capture that value. It's actually made way easier with good tax planning tools. Corby is my tippy top recommendation. Other tools in the space that do this step. Core V is actually the number one one that I recommend. Check out the link down in the show notes to learn more thought number four on what's to come in 2025 man receipt management document Management Couple things changing here. Dext just sold to Iris. Dext was kind of the granddaddy of all high volume receipt and document management processing platforms where you throw all your documents in there and it will match stuff back to the ledger. And it was the most powerful solution for a really long time. I think it's peaked by pretty much any account. When a company sells like that, that's not a good sign for great things to come for it in the future and they're going to say nothing's going to change. That's how it always goes with everyone, right? But I mean look at there's an unbelievable number of examples tsheets sure, prep when Thompson Road has acquired it like usually kind of hits pause and in fact Iris, same company acquired Senta, a practice management system, Iris has a competing practice management system, but Senta, the old platform, just a zombie platform now that's not getting any support. So Dex right now is probably still the best tool to do this job. I think you just probably won't expect as much development to continue there as you otherwise would have. But it's happening at the same time as the document management solutions that are built into all the other platforms are getting better. And if we circle clear back to the conversation around AR and AP and purchase orders and AI making that better than it's ever been before, it's gotten easier to do that stuff really, really well. And so when you see an uncap or a keeper or something like that, or Financial Sense or Kleinup Ship a version of receipt management that will do, and in some cases will even do like matching back to the ledger and push that receipt to, you know, QuickBooks or Xero, the argument for a standalone solution for document or receipt management is getting weaker by the year. I would, I would argue the last few years it's been a pretty tough argument to make depending on what you're paying for Dext because it is like a premium solution for that problem. Totally makes sense for businesses where you have a high volume of document management stuff happening. But that argument's been getting tougher because the receipt management functionality built into all the platforms, even the ledger platforms, has gotten better. Another thing happening here is increasingly receipt management is becoming part of spend management. So at the point of spend or putting something on a card, increasingly the receipt capture is happening there in a Distributed way across the workforce, which is a great thing, having to go back and track down receipts later and after the fact, like that's kind of the worst case scenario now where you're better off using a ramp or a relay or a Mercury or a row, where everybody on the team has these cards and they have a really good mobile app that goes with them. Where anytime something is charged on a card, depending on the spending policies, you may be required to upload a receipt and it may even disable your card after a week if you haven't uploaded all the receipts that you need to. And the notion that all that stuff is captured at the time of spend vastly simplifies the month end closing because you're not having to go down and track all that information. Right. Well, it is just yet one more place where receipts are probably better sent than some sort of separate standalone mega database of all your receipts. As intelligent as it can be. And I think Dex is the gold star here, I think the argument for standalone platforms like that are getting weaker by the day. And so document management, I think increasingly it's just becoming a feature, not a thing that you go out and shop for and actually buy. I mean you think about something like rent, that's free, how good receipt management is in ramp and the fact that you don't have to pay a thing for that, like that's, that's pretty compelling. Number five. And the last one, number six is more like crystal ball. Like what are some more off the wall predictions? But number five, I actually think the biggest change this year in running bookkeeping firms will come around the maturity of the firms themselves and the resources that you have to run a better firm. There is like, this is the golden age of free, really good, really high quality content out there to learn in years. What it has taken many practice owners decades to come round to like the level of thought leadership out there on running accounting firms. As much as we want to crap on like, oh, there's more accounting coaches than there are accounting firms these days and there's more communities than this and that's as much as we want to dunk on that. It's not a bad problem to have an abundance of resources. I don't think there's wrong ways to do that like paywalling everything. Like, I mean 99.9% of all the things that I ever do are out there for anybody to consume for free. And goodest takes for closing in on 400 episodes of this podcast that's just about talking shop about running accounting Firms. Gosh. You ever look in the mirror and think, oh, what have I done with my life? Then you see episode 400 take over and you're like, what if I become a doctor? Anyways, I digress. I don't. I have no regrets. But the bigger trend, I think is just. I don't want to say it's a generational thing, but the. I would argue the barrier has been lowered to have really high quality information and, and more support and launching your own firm and running a more profitable version of your firm. Like, if I think about, I think about the. Maybe the more legacy accountants that you came up with. Who, who were the big bosses when you were a young, fledgling accountant. And where those people learned it was usually like maybe a couple technical CPEs a year. Most of them didn't know a single other person that ran a firm besides just a passing glance and hello at church or at a networking meeting or something like that. These days, most of, maybe not most, a lot of you that listen to this, you have tens of friends who run accounting firms, which is amazing. The notion that you could text somebody with a question about this or that, have that level of access to other people to do this stuff is so good. A lot of you still don't have that. And you're lurking and you need to stop. Okay. Best thing I ever did was decide not to lurk. Six, seven years ago, I was not a social media guy, never posted on social media, was a 100% lurker. And look at me now, 400 episodes into a podcast. Maybe a cautionary tale, but you know what I mean, it pays to have friends that do this stuff and can give you some of the shortcuts. I also think social media is growing the market for specialized firms because social media gives you a path towards a path to micro niche. That's easier than ever before because I could talk about bonsai. Oh, gosh. We had this client, it was like a bonsai plant. You know, the little, like, I think it's a, I think it's a Japanese thing that, that is now everywhere but the little baby. Like the miniature tree sort of things. And they're very cool, like, amazing. But we had a client that did like bonsai tree implements and it was like a tiny hedge trimmer. Like it was, it was, it was all that. It was like all the full size things, but itty bitty, you know, let's say you're super into bonsai. It is now easier than ever before to build a specialized firm around that or, or around, you know, baseball card collectors and resellers and like whatever that interest is or that area that you want to go into Now. Social media has made it easier than ever before via algorithms to put content out that's for a specific type of person. It is not content that will do amazing vanity metrics or get 100 likes or anything like that. But you don't need to when you run a small firm like this. Like that's kind of the cool thing is you can put something out there that gets 200 views and all it takes is one person and that resonating with them and them being like, oh, this is a much more interesting thing that they're saying than what my generalist accountant says now. And I got a great new client. So I think there's a greater level of specialization and greater level of marketing sophistication coming from accounting firms than we've ever seen before because there's such an opportunity there. I get, I don't know, I go back and forth on marketing agencies. I see marketing agencies getting better as well. But the firms I get most excited about is like the founder who will take the time to post to social media every day and engage with them. And it's a little more of a. I'm obviously biased because I'm super into creator businesses and kind of what I do, building around thought leadership. But we're seeing in virtually every industry the idea of pairing a creator with the business and what an interesting way that is to acquire clients. And so even if you don't identify as a creator or an influencer or something like that, some version of your expertise and the things that you have to say living out there in the world serve to attract people to your business in a really exciting way. And I'm seeing that more and more in the wild. And I think that's really cool. All right, last one. Some kind of crystal balls, a little bit more like YOLO predictions. Things that may happen this year, maybe they're a few years out, but things that are kind of possibilities that could create larger shifts in how you run your bookkeeping practice. Number one, I think we may see a bigger volume of kind of Chrome extension based products to automate your ledger. So there are inherent limitations in the ways that QBO and Xero can be automated. And the biggest one is in the transaction matching screen. There's certain things that those platforms make available via API. So a basic example is bills. If I'm using Bill or some other AP platform to process accounts payable, that then uses QuickBooks's API to sync those bills back so that the bill looks the same in the account and the bill pay platform as it does in the accounting ledger. Right? Like that's an example of a really valuable, a really valuable part of QuickBooks API that they extend to other developers so that they can push information into the platform. But many parts of the platform are not available via the API. And for both QBO and Xero, a really core thing that's not available via API is the reconciliation, the matching experience that is, you know, rises theoretically enough to the level of the secret sauce of the platform to where they're not going to let other people use it. Arguably they're also doing their own kind of first party banking connections too. And if you could pull that stuff via the API, maybe there's an opportunity of for abuse there where you're like piggybacking that platform's banking integrations. Either way, the matching stuff can't be fully automated unless you have software that is actually automating the user interface itself. And so, you know, 2025 is supposed to be the year of the AI agent, that is AI tools that cannot just see the things that are on your screen but can interact with them. And we're already seeing the earliest versions of this. But the most natural way to do this is with a Chrome extension that's part of your browser that can interact with things and doesn't require like a big complicated thing running on your computer that can see absolutely everything. It is just sitting there as part of your Chrome browser and can interact with things within the window, you know, write tool. Hector Garcia's utility for QuickBooks Online is a Chrome extension that kind of improves upon the user interface and gives you some more power user features and even kind of an, a different menu. That is a great example of something that's become wildly successful, the app write tool. And let's link that down in the show notes that in my opinion rises to the level of like any account that uses QuickBooks online should be using WRCTool. Like it's, it's just that good. But it doesn't go to the level of like automating a huge amount of things in the interface. At least today it doesn't. And so just hypothetically, let's say you had a novel approach to being able to match transactions in an automated way or something like that, or do something more kind of across the accounting system, like more open ended, like maybe take a PDF of a bank statement and annotate it, check off each transaction while completing a bank reconciliation. Maybe that's a good example because it requires something in the platform, the bank reconciliation user interface and the PDF of the bank statement. So let's say you had a Chrome extension where you upload a bank statement and from that bank rec screen and QBO will complete the bank reconciliation and notify you of any transactions that are missing. Super useful, right? Well, the easiest way to do that today is with a Chrome extension. And so I could see more specialist Chrome extensions being built atop the apps that otherwise can't be automated quite in the way that we would like. Now, the reason historically you don't see this more is you're building on top of somebody else's property, like qbo. Theoretically. Like there's ways that you could make this harder to do, although with AI it's actually getting harder to kind of eliminate the ability of these third party tools to use their platform. But this is going to be a bigger conversation for these platforms in general. If we truly are getting to the age of AI agents, where you have all of these pieces of software using tools the same way that human users use them. Like I've said this quite a few times, there's a very high likelihood that 2024 was the last year that there are more human users of most tools than machine users of most tools, because it doesn't require a particularly capable machine to navigate a set of user interfaces. And if you ever want to have a play with this, take a screenshot of your screen, give it to ChatGPT and say, I want to print a balance sheet. What do I click next? It'll tell you reports I want to, blah, blah, blah. It's already good enough to do this. You can give very specific instructions, say, I need a cash basis comparative balance sheet for 1231, 2024, and it can navigate every single step of the UI of where it needs to go to do all of these things. And this is kind of the future of how we're probably going to be using computers as more assistance going out and doing things for us. And so the underlying platforms themselves, whether it's your ledger, whether it's your practice management system, they got to decide, are we actually cool with this? Like, this is actually going to put a probably a pretty big strain on our resources where you have more people using the platform than ever before. But a big percentage of those people aren't actually people, they're actually machines. So take a practice management system that does per user pricing, for example. Well, what if you have one user that uses the platform a thousand times more than any of the other users inside of a firm because they're doing all sorts of automated stuff on the platform. Is that okay or not? That kind of. The next level of thought that goes into that then is, well, do we actually go as far as making our platform easy for machines to use? Historically, platforms have always been designed in a way that was easily understood for a human. But is it also easier understood for. For the machine? Because there's maybe a market opportunity to be the best platform for the machine user, especially if that's where everything is going, right? So you kind of got some interesting conversations to be having there, which makes it more likely that these platforms have to be open to people automating on top of their platform. Because what's the alternative? Like if, you know, Xero doesn't lock their platform down for bots at all, but QBO makes it impossible for AI agents to use their platform, which even being able to stop that, I think has gotten incredibly hard. If there's value to you, the user, and being able to automate the platform, you're gonna go to the platform that will let you automate it naturally, right? Like, isn't that kind of inevitably where you go? So I think the market's getting better for platforms on top of platforms. And so we'll probably see more of that, especially around kind of legacy tools like QuickBooks Online or Xero that aren't themselves innovating that much anymore. This episode is sponsored in part by Financial Sense. If you were plugged into my 2025 practice management system recommendations, you would have seen Financial Sense on the list. I know a bunch of really smart firms that run Financial Sense. I have. I've been on record as saying it is arguably the best small firm all in one. Er, right now it's really hard to have a PM that does both tax and accounting work well. And while I'm not a believer in all in ones that do everything from tax intake to nailing proposals and all of that, what I am a fan of is the platforms that get closest to that. Right. Because at the end of the day, I do want all in one and I also want them to be investing in a more all in one path to where maybe right now that feature isn't good enough for me to replace that other tool, but I could see how two or three years down the road it is. I want that as opposed to the platforms that aren't investing in that way. Right. Financial Sense has actually absolutely been crushing their marketing the last couple of years. They run an annual thought leadership event for accountants. It's really good that I've been a part of. They're putting out actually really good social media thought leadership as well. Not just around like buy my product, but around better ways to run small firms. And those are the brands I want to get behind. Like those are the brands I want to work with. They got a smart vision for the right way to do it and the product's going to be a reflection of that. Learn more about Financial Sense Check out the link down in the show notes this episode is sponsored in part by Stanford Tax Gang. Stanford Tax was one of the like original, original originators of cool new tax workflow companies who have built killer stuff around intake. Specifically I've been shouting from the mountaintops tax firms heading into 2025, you need a tax intake tool. Will save you a massive amount of time. Stanford Taxes is one of only a few that I recommend. The whole having to like sit down and go through when you've got a stack of pile of client a pile of client information and you're like okay, here's everything we had last year. Here's what we need this year. What have they actually given us? Stanford Tax Gang. It will generate a complete checklist based on prior year pro forma information of everything that you need and you can drag all those documents on and it automatically checks off all the items. A stack of 40 PDFs, drag them on there, it'll auto resolve all those checklist items, will follow up with your clients until they get you all the stuff that you need and just completely eliminated that entire step of doing 1040 work where you're like do we start the project yet or do we follow back up with the the client for more information? Absolute game changer. Big fan of Stanford Tax. Check out the link down in the show notes. Another crystal ball thought here. I think we will probably see from the ledger players greater investments into month end close automation or even acquisitions of month end close automation providers. So when the general ledger is pretty mature and more or less what you know plenty for 99% of small businesses, what do you build next? In my opinion, and I've been telling the ledger platforms this for years, in my opinion, you build the next layer of abstraction on top of that. That is it's not enough anymore to be the best platform for a small business to run. I need the best platform to run a profitable accounting firm. And that looks a little different. That gets more into the life cycle of how I request information from clients how that information makes its way into the accounting ledger and then how I manage a large volume of closes at kind of a 10,000 foot view rather than the only option being inside of every single general ledger. Right. And so we've seen this from providers outside the ledger, like people building that stuff on top of the ledger. QuickBooks was actually very early to this with QBOA and some very light kind of month end close management functionality inside of QBOA that they shipped four or five years ago. You get a little bit of this and then they never really invested it in it anymore. Down the road upmarket you got stuff like FLO Cast, like there's entire month end close automation platforms. That is a much more normal thing up market. Probably because you have more stakeholders involved in that month end close and you got to get information from more places. But it's weirdly enough never really come down market to SMBs. So much like dedicated tooling for automating a bunch of month and closes. This is just a logical progression for me for these ledger platforms to have more tooling around managing the month end close or make acquisitions that will make them more of the preferred platform for not just doing one month end close, but doing 100 month end closes via a more standardized process. Crystal ball thought number three. Right now in the software sort of fundraising space there is some trepidation from investors to invest money into stuff in the accounting and tax space. Because I've heard this from a number of founders now because they're afraid of the near term impact that AI could have on accounting. And if AI just comes into accounting and automates a whole bunch of stuff, then it may not be a wise decision to invest in software today. Now I'd make the argument that I don't think that is imminent in what we do. I think there's actually a much greater perception of repetitive rote tasks in the accounting profession than there actually is in practice. Yes, we're all subject to the same rules and doing things similar ways. But what's unique in every single situation is the context is the client situation which impacts the correct answer. Making much of what we do not repetitive but actually unique. And one often requiring somebody's expertise to make a determination. But the tools in the space who are getting investment right now to do like a fully automated, you know, think about the Holy Grail goal kind of being a fully automated month end to end month end close. What would it take for AI agents to make all the requests that are necessary, all the determinations, review the books like clear from end to end, from gathering the information to closing the books. Without a human having to get involved in the process, Humans are probably going to be involved in gathering the information required to make those determinations. But that's no different than me sending a request to a client saying, I need XYZ in order to know the right answer to this. Right? That's going to, it's going to happen one day. Like almost inevitably you're going to have a computer that will do the close from end to end. But what makes that really, really challenging today, and I've talked with a number of startups and even folks on the investment side who are like, how real is this? What makes that hardest today is how different the tooling is and the inputs are from business to business. If every single business used the same five tools and had all the exact same inputs, we will probably already have this. The problem is businesses use different unique combinations of tools. They have different things that need to be tracked month over month. And because that is so different, even though supposedly gaap, which small businesses don't operate on gaap, but even though, you know, double entry accounting supposedly should be the same for everybody, it's just not like different companies have different needs. And so the notion of building a single platform that does all that stuff in a way that works for everybody maybe isn't realistic today. But where that starts, and this is kind of the point here, where that starts is a really powerful month and close automation solution for a specific vertical. So if one day robots will close the books completely from end to end, where will that start? Well, it has to start in a very specific vertical where they use the same tool set and have the same inputs. Right? Like that's, that's kind of what makes sense as a starting point. And so if one of these startups really nails it, I think that's how they do it. They go after construction or farmers or. And those are not even specific enough. Like it's a very specific type of construction that uses a specific set of tools. It's got to start very, very narrow and specific and then you can understand how then over time it can get more broad and AI getting better gives it more flexibility to be able to make determinations to make it even more broad then. But AI is not good enough today, so it has to be very specific. But if anybody nails like a fully automated end to end month end close, I think it probably happens in a very, very specific vertical. And to me, like if I were a founder in the space that was raising Money on the premise that we're going to fully automate an accounting month and close from end to end. That's the direction that I would be going, is find a specific vertical where we can actually nail it if they have a common set of tools and a common set of inputs. Last one, number four. You ready for this? Let's get weird. Large language models become the ledger, become the account exam. Excuse me. You can do my accounting in ChatGPT. Actually, hear me out here. In fact, I'll go track down this Twitter thread. It is something that I posted maybe 10 days after ChatGPT came out. And it was when everybody was token in the tires on ChatGPT and kept challenging the notion of like, oh, my gosh, can I do this? Kind of. Oh, my gosh, can I do this? Oh, it kind of can. I just ran a quick test and said, could ChatGPT be an accounting ledger? Let's begin. I'm going to contribute $100 in cash to the business. Now present my financials. And it's like, okay, here's the debit, here's the credit, here's your financials. And I was like, oh, that's. Well, that's. That's actually correct. What a return. Okay, now I'm going to buy some software for $50 cash. Uh, or you give it the dev and you give it the credit, whatever, and it gives you the journal entry and then returns with financial statements. It's like, oh, that's actually pretty good. And as I was running this test, I. I put it through a growing number of transactions. And then I said, now every time I have a purchase from Office Depot, automatically put it in Office Expense. And it's like, got it. Anytime we have that come through, we'll automatically put it there. So then I give it three more transactions and it identifies, well, this one's, you know, Office Depot. We're going to automatically put that in Office Exposure. I don't have to ask you for it. And it was like, really, really good. There's another example I did where I went through this exercise of doing a number of transactions. And then I told after the fact, I was like, analyze all of the transactions we've gone through so far and automatically create rules according to how I've classified these things based on the vendor. Anytime you see the vendor in an amount within 200%, automatically classify it to that expense account. Pretty much what we do with bank rules. But the bank rules, I'm just telling it, make bank rules for these things. And it does. And it did it. And when I continued to give it transactions, it processed that stuff correctly. Like, and this was the very first version of ChatGPT. Now, if you want to get, like, super technical and nerdy here, there are constraints with this one being context limits. How much can it remember? Well, eventually it starts forgetting the earliest parts of the conversation, right? Well, so does your accounting. Like, when you go to your accounting ledger and you say, go, make this entry, what it does is it commits an entry to the database, and that is now in the database, but it doesn't remember. I guess you could have an audit trail, like, all of the inputs that you make. But you can do the same thing with ChatGPT. Like, you can have a database that lives behind that. And this is from a technical standpoint, this is called function calling. You can build function calling into the AI system where when it detects something, like, you know, let's say you use function calling to get the weather forecast for where you live. And I could be talking with it about my favorite books, but when something comes up along the lines of like, hey, what should I wear tomorrow? Or what should I pack for this trip? Then the language model goes, aha. We have a function specifically for this. It makes an API call to a more traditional service that has the weather forecast and then gives that back to ChatGPT, and that then presents it in a conversational format. Well, we can do this with an accounting ledger. We can do this with a database. Anytime it detects a new transaction, it could push that entry to a database the exact same way that your accounting ledger does, in a more structured way. In fact, we talked about this a bit yesterday in the tax episode, Microsoft CEO Satya Nadella. He says that language models are ultimately in the process of replacing software. In fact, you may have heard people say this. You know, software is eating the world kind of in the last couple of decades now, people are saying AI is eating software. The language model is kind of becoming the software, and then behind that, you just have a database. So you. You interact with a large language model, and the large language model just parks stuff in a database that it's going to need longer term. But the software kind of starts going away and you just have an interaction with the AI assistant. And this can 100% be a thing with an accounting ledger. Like it is, it is posting those entries to a, to a database. It can then retrieve, you know, balances and generate reports and all that. You can even generate invoices. We've done some examples of that with ChatGPT where you tell it the format that you want the invoice and it can generate the PDF and honestly it looks just fine. This actually wouldn't super surprise me if we got a decent AI assistant version of an accounting ledger, as wild as that sounds. Because another trend that's happening in the space right now is all the ledgers are going up market just getting more expensive every single year. I mean think about the minimum cost of a QBO ledger these days. They did ship quote unquote ledger, which is great. It's a low cost, super bare bones version of qbo kind of built around annual tax break and that's great. Xero's had a version of that for decades, for 10, 15 years where it's like nine bucks a month or something like that for QuickBooks Ledger, that's great. But these are very mature platforms and they're only ever gonna get more expensive. They're not interested in coming down market. If you are a sophomore in College right now QuickBooks Desktop's no longer available. Where do you go to use a free accounting ledger or a super low cost accounting ledger like and these, these entry level like low cost versions like QuickBooks Ledger are so stripped back, they don't have, I think, I don't think they have stuff like invoicing and a lot of stuff that you may need. And so unless you're willing to spend 60 bucks a month on an accounting platform, it's very hard to run a SMB set of books. Like where do you even put that These days Everybody has these LLCs with like a single rental property in it or you know, we have a whole bunch of examples of this stuff where there's maybe one core operating company that could justify a full monthly subscription, but all the other ones are like, they really don't need that. So all these tools are going up market, nothing is filling that void down market. But you know what could chat freaking GPT and it wouldn't actually be ChatGPT. It'll be a company that comes in and builds the function calling and they have the databases. But you are basically interacting with an AI assistant, giving it the bank statements, you know, giving it the CSV of transactions, having a back and forth with it about how it wants, how you want us to classify things. That may very well be the future of accounting ledgers. And a very powerful thing to remember, I can't remember who said this is Paul Graham or somebody else. They said you've already seen the future. You just dismissed it because it looked like a Toy. And that's how the future of everything starts is right. Now that, that's just a very cute idea, isn't it? Right, like, okay, well, my grown up clients, they all have proper accounting ledgers. I see that. I can see how, okay, yes, maybe you can do a version of this that is in some AI assistant form. And, and that's cute. Maybe that works for very specific applications. But we have our grownup ledgers over here. That's how it starts. As always, how it starts. And then like before long that thing gets more and more powerful and similar to the example of enabling dark mode in Windows, you're like, oh, I actually see now that it's much easier to just tell it to send Jim an invoice for $100 worth of apples. And it just does it. Rather than all of the clicks and all the inputs that requires. I can literally hold down a button on my phone and in six seconds give it the information that it needs to go and do that thing. Go do that job. Isn't that almost certainly the future of an accounting ledger? And when we're talking about this situation where QuickBooks Online and Xero have such market and network advantages because of the ecosystem, and why would I go use Puzzle or Campfire, some other ledger when it does more or less the same thing in a slightly better way, but doesn't have all the other tools that'll connect with it? Right. I think that Nets Ledger might just be an AI assistant and that is an example of AI eating software. There you go. My most off the wall crystal ball. Seemed like this might be a thing. Maybe it won't be, but if it did, you could see how we get there and how it actually fundamentally changes how accounting gets done. Right? Lots, lots of ways you can shoot holes in that, but also ways where like that's something you could go out and build tomorrow. And it's actually kind of interesting and somewhat viable at a very, very basic level, which is how everything starts. So there you have it. That's, that's my most kind of off the wall prediction for 2025 or maybe the near future after that. But I'm curious if you've actually made it to the end here. What are the things you're most excited about? What would you like to see? Believe it or not, there's going to be a ton of software people that listen to this and pull all sorts of ideas out of these podcasts. And that's a big reason why we have these conversations, is every time, every time we're talking about what's possible with AI. And when you bring this stuff to your vendors, it's raising the quality of conversation that we have with software folks in this space, which I'm 100% for. We kind of become advocates for a better. Better tools for. For professionals like that enable us to do more than ever before. Because, gang, if we don't have really good tools in the future, that puts us. Puts our profession in not a great spot. And I wouldn't argue that our profession has leaned into AI solutions as as much as we could have, as evidenced by a lot of the incumbents really not shipping meaningful AI stuff and not having an understanding of, of how powerful this stuff is yet. But the more you have these conversations, the more I think we kind of raise the bar. Raise the bar for everyone. So if there's anything that jumped out here where you're like, boy, I would love to see this or that in this tool I already use, send them a clip, send them a time index of the YouTube video and say like, hey, are you guys working on this? Because this seems exciting to me. And there's also a degree of manifestation here. Like we just had a conversation about this, you know, kind of general ledger that is, that is, you know, just an AI assistant like this. These conversations will also inspire people to go out and work on these problems. And they may go somewhere, they may go nowhere, but at least we're progressing the conversation by being here and engaging with it. Anything I miss, anything else you want to see? Drop that stuff in the comments. And here's to a killer2025 game. Let's do it.