419 The State of Accounting AI Adoption

It's that time of the year, baby. The annual State of Accounting AI report from Carbon just released this week. It's got a lot of fun tidbits in it. It's got a, got some spicy quotes, got some good ideas, some stuff I don't agree with. We're gonna be going through the best of it today. Come on in, let's geek out. Coming to you again today from Mealman Australia, talking all about this, this Carbon report, I should say because I'm a, I'm a man of integrity. This is in no way sponsored by Carbon. I just think the report's interesting. I actually think there's a lot of really good tidbits to take from it. I also want to fight with some of the people that provided quotes for it and that that just makes for good content. So here we are. Ironically enough, I think Karma might actually be a sponsor on this episode. But I'll be darned if I'm gonna let any, any corporate software platform tell me what to talk about on my podcast. Okay, we'll put a link to this one in the show. Notes. Why does this matter? Because this is, I mean this is a, a once in a career time, not lifetime, but career time shift and how work gets done. It really truly is. And if you don't realize that yet, take my hand. We're going to walk through some of why that is today. But in short, it boils down to the emergence of intelligence as a new resource, as a new oil. So oil, you've heard of it, it's used in all sorts of applications. It's a multi purpose fuel. What you do is you take some dead fish, you give it a couple million years and it turns into a flu that you can put in your car. Now, AI being like this sort of multi purpose knowledge. It's the idea that these large language models can come to intelligent decisions, either quick, simple decisions like is this right or is that right? Take, you know, a, say a bill pay scenario, look at everything in this bill and here's the last bill that we paid. Does it look like there's any duplicate, you know, stuff in here? Does it reflect the fact that the last bill was paid? Something like that. That has to be a little bit open ended where there has to be like a little bit of intelligence applied there to make a decision. That sort of thing isn't made possible by say OCR just extracting information. It's not possible to do that in a highly structured way that a few lines of code can handle. It is intelligence In a very, very basic form that today we're beginning to use in like these sort of chat assistant contexts. But increasingly we'll just make decisions where humans don't need to be involved. And if I have an objection with some of the quotes in this report, it's that, it's that our minds are still telling us, oh no, AI is going to come for the menial work, buddy. AI is coming for the entire lunch, not just the first course. There really isn't anything in the stack of things that accountants do right now that will not be impacted in some way by AI. And that's hard for us to get our heads around right now because every tech change in the past, it's been the big boss talking about how it impacts the junior folks. Meanwhile, you have the open AIs of the world, the people building these models who are saying we're a couple generations of models away from models forming organizations and the models orchestrating what's happening in an organization. Do you want to get a little bit sci fi here? Humanity, we like to think of human knowledge, human capability as the ceiling. Because as it exists right now on our planet, we are the ceiling. We are the top dog, is the great place to be top of the food chain unless you end up cornered in an alley by a polar bear. We're it. Like we've got Earth figured out and so we can't really think or consider what knowledge beyond the knowledge of humanity looks like. We could also take a goldfish and you could take the average sort of intelligence of a goldfish and you could, you could give that a score. If humanity is 100, your goldfish, it's you know, five or something like that. Now AI, specifically generative AI, where we can ask it to do these open ended things, few shot, zero shot, you know, you ask it how to do a thing without even telling it how to do that thing. Before generative AI, which is all of three or four years old now, its intelligence has been increasing rapidly, right? So if your goldfish is five, you can see how the very first ChatGPT may have been, you know, 25 or something like that. And you know, intelligence is not this linear. It's much better at some things than other things. But, but I think you get what I'm kind of getting at. And it has been getting more intelligent at a shocking rate. And a really, really big unknown is if humans are 100, will AI stop at 100 or will AI stop at say 110? That is a reflection of the smartest humans or we'll just keep on going because there's nothing to make us think that humans are some like absolute force in the universe where nothing can be smarter than us. Now the argument for why it may stall out at human level intelligence is that it's trained on human stuff. So how is it going to be smarter than humans? The, the folks with a lot of money to be made by AI being very smart will tell you no, they'll go off. These AIs will go off on their own, they'll talk with each other, they'll work really hard and they'll come up with novel things and we'll have sort of this compounding effect where they continue to get smarter and smarter well beyond the plane of human intelligence. And if this is the case, who you think is going to be working for who? The big boss, the one that's been protected from all the other automations and tech changes in the past. The big boss is going to be working for the AI just like everybody else. Now everybody has different reactions to this where it's like, well we got to put the genie back in the bottle. This is bad. We can't have anything that's smarter than us. There's other people that are like, this is like the crowning achievement of humanity to have this thing to make a thing that's even better than we are. It's the difference between the dad who when the family gets back together at Thanksgiving, wrestles his son in the yard, his 45 year old son in the yard just to prove he's still the top dog, and the one who wants to see them to go on and do something better than they could have ever done. That's a great analogy for humanity's relationship with AI right now. So if you think you're safe because you're the big boss, let me tell you, there's a whole bunch of different ways that this could go where we're all totally safe or it's gonna come and help with everything that we do. I can tell you a really, really not helpful narrative is that AI will never replace you. What I find it does is actually encourage this level of complacence or it enables this level of complacence where it's like, nah, I'll be fine. The reality is everything's changing. It feels like it's going to change fast, but it's not as if we haven't had change before. I mean there's a time when 98% of us were farmers. You know how to farm. I can tell you I should know my wife is going to run the fine folks at Lowe's out of business with their kill your plant and come back and get another one for free guarantee. My household is so far from being able to farm, but when 98% of people were farming and then, you know, industrial revolution comes around and everybody says we don't need to farm anymore, it was really hard for those folks to see over the next hill to see like, what, what are we going to do next? Humanities it has always seemed to find another thing to work on. There is data like over the course of the centuries and millennia, especially the last few centuries, that say we are humanity is working less than ever before. It was not that long ago where all you did was work, like work was just survival. And now that's definitely not the case. But I mean, when you're out in the potato fields, your outlook notifications weren't going off on your phone. So don't act like you understand what we're going through. Things are definitely going to change. Honestly, incredibly hard to predict what that looks like. I can tell you right now there's a whole lot of people that really still need you. A whole whole lot of people. And having an AI that's capable of PhD level stuff is not the same as an AI that can go off and self govern and help everyone. Because right now the AIs that are solving PhD level stuff, you know who's getting them to do that? PhDs, there's a reason that you and I aren't getting AIs to work through PhD level problems. Let's be real. That's just, that's not us. Far from PhDs. You're trying not to fall off an exercise bike right now. I'm wearing Croc sandals so time will tell what happens. But the what we all have to do is like lean into AI and educate ourselves on AI. Don't let our teams get left behind. I said this, I think it was last week. I'll say it again. You think about the, the percentile that you are in in this profession. If you listen to this podcast, you take the profession as a whole and what percentage of the profession is turning up online and learning from other practitioners online actually thinking outside of their firm's group think bubble. I would, I would estimate less than 5%. Most people are just heads down playing the game that's been defined for them inside of their little firm's universe. Of that 5%, how many are like actively engaging in learning content like this? Probably 5% of 5%. And I'm very biased. But I think the group that's tuning into my content is probably the best prepared accountants globally for what's happening in an age of AI. And a lot of stuff's going to change, but there's also going to be people that win in that age. And if there's somebody that's in the driver's seat for it, it's all of you who are engaging and staying plugged into this stuff and actually putting AI to work. Okay, let's get into the report. Some of my favorite stuff again. We'll link it down in the show notes. Basically, Carbon goes out and grills a bunch of their users on their sentiment on AI. And the interesting thing they're able to do is see how that changes and evolves year over year. Some stuff has gone the direction that I expected. A few things haven't. One thing I like is most of the folks surveyed in this study are small firm folks. So over 80% of the people surveyed are in firms. 50 people and under. The strongest representation is among firms between four and 10 staff. So these folks, they look a lot like y' all that tune into this show. Half of the respondents are partner owner type folks, while the other half are the folks doing the work. That was the wrong framing. The other half are non partner owner people. In terms of folks excitement level on AI, they ask everyone a question that asks them to sort of self categorize in one of four buckets. One, they're excited about it. Two, they're intrigued. Three, they're skeptical or scared. Four, they're indifferent. We should also note these are Carbon users, so these are actually a very progressive subset of the profession as a whole. So these numbers will definitely skew more AI positive, I suspect, than the general accounting universe. This episode is sponsored in part by Carbon, the practice management system. Something that Carbon does that I love is they will literally tell you the stuff that they are working on and have like a big TADA event at the end of the year to announce. Like, hey, here's the stuff we're about to ship and here's what we're prioritizing for the following year, which that level of confidence and transparency I I like. In fact, in the process of doing that, they probably actually get a lot of really useful feedback around. Like okay, here's what I like about that, here's what I don't talk about that. Whereas if they didn't put that stuff out ahead of time and kind of just did this big ta da, they would have less information from their users than they have by like kind of involving them in that conversation. They did that at the end of 2024. Announced a whole bunch of stuff. Honestly, all the Carbon nerds. And Carbon just has some real enthusiasts. They all turn up to that webinar and they're so it's like the highlight of the year. They got the popcorn out there, like, oh, what are they going to announce? And a bunch of people did get excited and they went ran out to LinkedIn and they're like, oh, I'm so, I'm so excited for this and that and that and that. I mean, props to Carbo for building that level of fandom in your software tool. But also people, it's accounting firm software. Let's get some perspective. Probably not a lot of tools in your stack with that level of transparency if you're in the market for practice management system Carbon, it's one of those tools you kind of have to consider. It's like on the have to look at list. Learn more about Carbon. Check out the link down in the show notes. This episode of sponsored in part by Teamup. You know, the greatest proof of something that actually works. People are doing it. People are coming back and doing it more. I got a report here from the last few weeks of the projects that Team Up's been working on. By the way, Team up, they help you hire talented accountants in the Philippines directly. You work directly with them. Can I just run you through like, I don't know, like half of the things they got in the pipeline right now? Firm in Hawaii, hiring a bookkeeper. Returning firm. They're coming back for more. New firm in California. Hiring a senior accountant. New firm in Australia. I'm so sorry. Oh, what was that? I'm so sorry. Hiring a tax accountant. Firms in Florida and Texas hiring senior tax accountants. They are returning firms. BC a returning firm from BC hiring a junior bookkeeper. Arkansas firm Hiring an executive assistant. New firm from Colorado hiring a senior accountant. You get the idea. Every flavor of accountant you can imagine. A whole bunch of firms that are coming back, which tells me they've done this before and it worked. That's what I did. I started with one and then immediately hired like three more folks in the Philippines over the course of two or three months. If you're looking to work directly with folks in the Philippines, Team Up's got the recruiters that can help you do it. Check out the link in the show notes to learn more. This may come as no surprise. Leaders are more excited about AI than individual contributors, non owners. So 63% of firm owners report they are excited. That's like the peak level of stimulation, followed by 25% that are intrigued, which is slightly less stimulated. That amounts to almost 90% of. Of firm owners, individual contributors. On the other hand, 40% are excited. So that's over 20% less than firm owners are at that peak level of titillation. Interestingly, this has increased. The level of folks who are excited has gone up by about a third since last year. So it has gotten more real. The number of people that are saying this is just crypto has gone down. I do think the leadership skew is kind of just a product of what you consume. So leaders consume stuff from other leaders. They're more likely to be outside the thought bubble of their own accounting firm than individual contributors who are more focused on playing the game that's defined to them by their leaders. And I think this is why across the board, you see the C suite, the executive type folks are the ones really, really pushing AI. But the flip side of that is the people who are doing most of the, the grunt work, the heavy lifting, they're like, you don't have an understanding of what I do to the level that I do. So who are you to say that this is the answer for my job when I'm the one that does my job? A quote. There's a good number of quotes in here that I'm going after. I'm not going to say who any of them are from because it's not really about their commentary. I also think the way these reports come together, like oftentimes this quote can be pulled from a whole bunch of stuff or a longer conversation. So more just kind of digging into the sentiment here. But here's a quote. To me, AI is a powerful tool that can augment human potential, offering new ways to understand the world, solve complex problems, and improve lives. I hope so. I sure hope that's right. For some reason, we feel the need to say it's going to be our co pilot, it's going to be our buddy, it's going to be our best friend. It will, but it'll also just do stuff. And that's okay to say aloud. We're lying to ourselves if we don't acknowledge the fact that it's just going to go do stuff that we used to do and we'll have to find something else to do. I think there's probably plenty of stuff to do, but we don't have to be so attached to the notion that this is our companion. There's such a emphasis on a cozy narrative, but the fact is it's just going to do a lot of stuff that we've had to do before. Just some of the stuff I've been sharing on social media this week around how good research tools are. Now for tax research. OpenAI's deep research perplexity Zone Deep research implementation, which is free. You can do five deep research searches per week. If you haven't played it, go to Perplexity AI create an account. You don't have to pay anything. Just know that Perplexity trains on your data, so don't put any client like identifying information in there. But it is the best tax research tool we have ever had and Perplexity's version is free. The notion that in two to three minutes I can get really, really high quality answers and the entire chain of thought for how it got there, how it kind of has conflicts with itself and what makes it go out and seek more information to clarify any gray areas. It is so clear. It's so much better than just an answer because I can see that thought pattern and be like, does this make sense for me? Like, would I have gone a different direction at this point? That is incredibly useful because even if I don't agree with the final answer, I can at least get 50% through its logic and then like, choose my own adventure. But in my experience, it's incredibly accurate as long as it can find the answers publicly, available online or in documentation that I provide it with. Now, how was that, how was that done? When I was a junior in an accounting firm, you cut somebody loose and they would spend a few hours like gathering the most relevant materials. And that can be time consuming because tax research involves finding all the different places that give you an answer and hoping that they agree, but also like reading the entire rest of the thing because six pages down the line there's going to be an exception to that thing that you thought was the rule. And that can be really frustrating. And this is where AI being able to pull across a huge number of sources, both authoritative and like commentary from experts and present the most relevant stuff to you is a tremendous time savings. That's work humans were doing that we don't have to do anymore. Is it a great companion to the senior or whoever has to make that decision? 100%. But let's not get ourselves. It's gonna do work that we were doing and that's okay. We gotta trust that over, over that hill there's going to be plenty more to do. And right now, I mean there's, I mean just think how little like time we spend with our clients, how little of the human stuff we do because we are spending so much time on the technical, on just the getting projects out the door, on the managing of the work itself. Sure seems like for the foreseeable future there will still be more valuable things that we can work on. Another quote here. AI in the short term is a nice additional tool in our toolbox. As accountants, I believe over the next one to three years, AI will be built into every tool we use. In the mid to long term, it's anyone's guess whether AI causes more issues than it solves. Yep, that's a pretty balanced statement on the subject of it being built into the tools that you use. Try this one on for size. I think there's actually two potential timelines here. Either AI gets built into every tool, as this person suggests. What's hard about this is it takes time and it means overcoming infrastructure like what you've built to date. You've got this big machine and shipping updates to it takes a long time. There's people and inertia and all that. And implementing AI into it is tricky. And if you've seen the volume of AI updates just the last two months, from all the deep research stuff to operator like automating browsers, to new models that are capable of things that old models could have never dreamed of doing, keeping up with the speed of all that in a mature product, that's really tricky. So two timelines. One, AI is embedded in the tools that you use. Or two, the alternate timeline. This is the wild one. AI becomes the tool because, you know, ChatGPT shipped about a month ago. Tasks. What? Managing tasks in ChatGPT. Recurring tasks even. What do you mean? How's that? This is like a to do app now it's starting that way so you can tell it remind me to do XYZ or every Monday go out and pull the price of Bitcoin at 8am Pacific or something like that. And right now it's kind of table sticks. It's useful for pulling time sensitive things from the web. It already is. There's a couple of use cases accountants have shared where they have to go pull Flynn values and these different things on a recurring basis. And using it for something like that is really, really cool because then when it does it, it can add it to a CSV or something like that. But start thinking about what else OpenAI has shipped lately. So operator only available on the $200 a month pro plan. But it has its own Internet browser and can navigate and autonomously do things on the web for you, do your bidding on the web. I mean, we are, we are real close to a reality where your task management in ChatGPT is tied to actual tasks being done for you. The agent going out and doing those things, it's not good enough for that yet, but that feels pretty close. So AI might get built into the tools that we already use, or AI might become the tool. Because if that agent's job is just to log into this platform and log into that platform and move the data from here and move the data from there, then it's just a matter of time before the AI is just like, okay, I understand what those platforms do, I don't need it anymore. I will just be all of those platforms for you. So you have this sort of, you know, just in time, real time software just kind of being managed for you by this large language model. Maybe it's backed up by a database, but I mean, there's a very possible reality where the AI you communicate with just manages all that stuff for you. Because if you think about, you know, let's pick on Carbon because this is their report. If you think about what a practice management system does for you, like it manages the database of all the projects that are in process and all the attached things like time of billing and invoicing, interacting with your clients, are these not all things that a large language model could do in the very near future? Now maybe we ultimately need that interface and that structure and that is, that is something that is core to keeping organized. Honestly, I don't know and folks at Carbon don't know. And Carbon's got some real smart like AI folks. So like, this is not a surprise that anyone is not like exploring and trying to get their head around. You think it's a weird time to be accountants, man? Imagine building software right now when you've just got no idea what software will even look like 24 months from now. But normally a startup is working on sort of this, how do we get to 100 million in 10 years? Or something like that. Who in the heck knows what the world's even going to look like in three years. I'll be honest, I would rather be doing accounting than software right now. But it is a very exciting time for both. Some other data here in terms of who's the most skeptical versus excited wee bitty firms, one to three staff, they have the highest percentage of skeptical, indifferent responders over 25% in total of one to three staff firms said they're either skeptical or indifferent to all things AI. That was. Oh, that's the 2024 number. 2025, that's actually cut in half. So we still got some holdouts. The people that say it's the same as crypto, but that's now been cut in half, like 13% total from the small firm owners. The other end of the spectrum, firms with 200 plus staff, they are the most bullish, which is actually really interesting that this tech change is so universally being leaned into by large organizations. I was listening to something the other day that was kind of going back to the cloud transition. And back in the cloud transition, folks were split. I mean there were a ton of organizations who were like, the cloud is never going to be here to stay. It will always be better, cheaper, more secure, all these things to run your own servers and all that. And there's a lot of organizations that were going to live or die on either side of that fence. AI, obviously all the AI companies are hyping it up to no end. And the VCs that have hundreds of millions, billions of dollars into these companies, they're definitely pushing the narrative that AI is the future of everything. But you know who else is like everybody else, whether you're an AI company or not, like everybody is drinking Kool Aid. And that's exceptionally rare. I don't know any other examples of something that folks have been so universally behind besides returning to office. Just kidding. In terms of how people are using AI, number one use case is the same as last year, communication. 63% of people are reporting they're using it to compose emails and fine tune their writing tone. This is because this is a feature that's built into Carbon, so that's an easy one. The fact that it's built into the app that they're already using, then yes, they're going to say that they use AI for this. More interesting to me, 40% of folks reported using AI meeting recorders for the first time. That is not a thing that Carbon does or any practice management system right now because it means having to have access to like your meeting platform. About 40% of folks report using meeting transcription tools. That's up 12% from last year. You love to see it. I know, I heard from a lot of folks when we did some content a month or so ago that they started capturing their meetings. That's great. We talked about this on channel for a long time that you'll never regret having those recordings 39% of people reported using it for research. But if you think about when this information was gathered, I suspect it was late last year. And boy, just in the last 45 days, the entire AI research landscape has changed so much. I suspect the percentage of folks using it for research is going to shoot way up. Another quote to tear apart here. AI is the tool we will use to complete the mundane, boring, repetitive monkey tasks. I hope anyone who left or gave any of these quotes are listening. We'll use it to complete the mundane, boring, repetitive monkey tasks, which will in turn allow us to focus on more complex and fulfilling tasks. Human relationships might even be positively impacted by this as there will be more time for human interaction. This episode is sponsored in part by Cloud Accountant Staffing, who helps you find super talented accountants in the Philippines. Boy, is it's 2025 going to be the year of the Philippines. Like, it's not even just an accounting thing. I have talked to like so many people in other industries lately where they're like, yeah, no, we've started picking up people in the Philippines. Just the whole trend of like folks in the Philippines working for Western companies that like, has become so normalized now. I was talking with a listener podcast the other day, completely unprompted. They're like, yeah, no, this was a good idea. I changed this and that and I kind of finally got off my butt and started hiring people in the Philippines. And I hired like three people with Cloud Accountant staffing just last month. Holy smoke. First of all, chill out guy. Second of all, what took you so long? Third of all, what's taking you so long? Okay, it's the same thing I did. I made one hire and I'm like, okay, this is actually working really well. I'm going to make some more hires. This is not me just shilling wares for people that do placements. Actually, I guess that's technically exactly what this is, but it's also exactly what I did. It worked really well. A lot of listeners who have started doing this, I've heard super, super positive things from they're still human beings, it's still managing people. So it still sucks in many ways. But Cloud account staffing is going to make it as easy as possible for you. You can explore that. Check out the link to learn more about Cast down in the Show Notes. Hey, real quick. I write a weekly newsletter, goes out every single Thursday and let me tell you, it is a banger. You ever struggle to keep up with like a high news updates? Things are happening so fast. Jason where do I learn what is the most up to date? Tippy, Top hot goss in the industry and tech advice and all that. It's the newsletter buddy. Straight to your inbox Every single Thursday, 2pm Eastern. Thousands and thousands and thousands of accountants read it. I'll put a link to it down in the show notes. If you aren't on that list yet, you better get on it. Most people, when they automate a thing or they're able to do 20% more of that thing, we always say so that we can get back to the more important things. But what people actually do is just more of the same. So there's a very real possibility here that AI makes us twice as efficient in the work that we do and we just end up doing two times the exact same type of work. That's the most likely scenario. That's what most people will do because that's human nature. It's what we do now. It's why when you're going through change management and you're thinking or talking about, I think with this change we can get x percent of time back. We need to have a plan for that time before it just gets gobbled up by the default, which is more of the same. That could be a more valuable service line. It could be any one of a number of things. We just got to look out for that because that's what we all default into. But also we're coming back again to the mundane, repetitive tasks. This is not every other type of automation we've had in the past that started at the bottom. AI is equally useful for senior folks right now as it is junior folks. Quote here from Chad Davis, Canada's own multi time guest on the POD part of the three Timers Club. He says AI to me it's more than just automation. It's a partner in reshaping how I work and connect with clients. AI enables me to streamline complex processes, saving time and boosting efficiency. Did AI write this? But beyond the technical gains, it allows me to focus on on what truly matters. Building relationships, sharing knowledge and helping clients see the future possibilities for their businesses. I'll be honest, this has got a bit of the bit of the same vibes of that last quote, allowing me to focus on what truly matters. It can, but that's not the default for people. If I make you 20% more efficient, the vast majority of people are going to go back to doing 20% more of the exact same thing. And that's okay. We just got to be aware of how that is like in our nature to just do more of the same, your team members are going to be the exact same way. Unless we very intentionally direct ourselves to something different, we're going to do more of the same. Respondents reported that they thought bookkeeping would be most disrupted by AI. And I'm just going to give you two thumbs down on that one. That's just not the case. This is another carryover of the notion that the, the lower value, should I say, or easier tasks are the ones that will go away first. When AI, it's not starting at the bottom and coming up, it's in a jagged way coming for anything and everything. Bookkeeping. And this is the biggest thing that AI software companies have wrong right now. They're like, oh, yeah, now we can finally automate bookkeeping. With AI, the time that professional accountants spend on bookkeeping, it is not spent on classifications. Bookkeepers, I guess, depends on how you define bookkeepers versus accountants. Let's talk about the month end close, though, because if as accountants are getting involved in bookkeeping, it is more at the controllership level over saying month end closes the month end close. The percentage of time that you spend on classifications is minuscule. But people think because AI can help push us from 90% to 95% automated classification, the, the AI is going to automate bookkeeping. Man, the month end close is all the other crap. It's tracking down people for information. It's why was this different this time? It's what the heck is this transaction that doesn't belong here? It's why did this payroll liability not get paid? Well, it was because Janice went on vacation and she forgot or they got locked out of the account to pay that employee's garnishment. It's a whole bunch of stupid stuff like that. You have so many software companies that are, they're trying to solve 200 IQ problems only to find out that we're all blocked by 60 IQ problems. It's just a bunch of human stuff and it looks a little bit different every single month. That's why the month end close hasn't been fully automated. Am I excited for more automated classifications? 100%, man. Like, there's no reason why in 2025 anybody needs to be making bank rules. Why am I manually creating rules for how transactions can be classified when I could copy and paste that stuff into ChatGPT and it sees it two or three times and it will always classify the same way again. Ledger platforms need to adapt. Rules need to be basically automatic. Like, don't make me manually Create these static rules and the better version of that, the next version of that, it'll be great. But it's far from automating the entire month end close. Right? The stuff that I think is more immediately disrupted by AI accounts payable. This has already happened. Accounts payable is absolutely getting smoked right now because AI vision models are so good. And then you use large language models that have a bit of intelligence to be able to make decisions in fringe scenarios. That is the most significantly displaced part of anything that falls into the accounting profession bucket. So stuff that's really going to go away quickly, anything that AI vision models can help with. So stuff that's reliant upon extraction or making sense of the layout of things in a document because it's light years ahead of ocr. And also anything that just requires a bit of intelligence because the language models we have right now, like they're not super big brain, but they can do stuff that OCR can't do. They can do stuff that a few lines of code can't do. And I don't see the entirety of bookkeeping falling into that. Another quote here. You might notice the theme, AI is a great tool, but doesn't completely replace a human. Poppycock. It sure does. If that human's job is to sit there and process accounts payable bills, you better believe it does. Now most accountants, they're doing a very diverse set of tasks. You know, somewhere in an enterprise they got an entire floor of people doing this one very menial thing. Those people get replaced 100%. But public accountants, by and large, you do a whole bunch of client facing stuff. You're doing a hugely diverse set of tasks, extremely unlikely something comes in and does all of those for you. But why, why do we feel the need to tell ourselves this AI is a great tool, but doesn't completely replace a human. That doesn't help me get any closer to how do I use AI. If anything, it encourages complacency. And this is the most important time to not be that. This is where we need to be thinking outside the box, applying new technology to our domain in a way that AI experts can't. Because if we don't, who will? You want some AI bro coming in and telling you how to do your job? No, this is where we get stuck on like not identifying as AI experts. You don't have to be, you don't like, you may not be an AI expert, but if you don't start applying AI to this stuff, who will? AI experts that don't know anything about accounting. You want those people serving your clients? Please, no. 70% of people reported they were concerned about security. That's down 6% from last year. It's gone down from 76% to 70%. That's a good thing. We need to be mindful about AI for sure. But there's completely secure ways to use AI. We need to continue despookifying it. Here's a quote. If confidentiality and data security was insured, then our firm would be using AI. Buddy. It has been. There is nothing spooky or unknown about it. Whether you're in your Microsoft ecosystem or Google ecosystem, whatever the AI assistant, the security on that is no different than anything else you have in those Systems. Talk about ChatGPT. The team plan a lot. Very secure fine to use with client sensitive stuff. Gotta keep shouting this from the mountaintops because it is blocking people from adopting now last one. And the biggest takeaway from this report is the difference between the people who report being excited versus the percentage of people who are actively training their teams on it. Over 50% of people are reporting they are excited. The peak level of excitement of all of the options, their maximum excitement. But over 50% of those same people are offering no AI training to their teams. Zatron Bones, man, that's a bummer. Why? Why is that? Does it not exist? And so I start thinking about like, well gee, how do we, how do I make some easy training content that everybody, everybody can go use? And I'm like, oh no, I already do. I make YouTube videos. And hopefully that gets people excited. But how do we solve this? How do we fix this? Everybody's so excited, but we're not actually taking the time to train our teams and get them comfortable with this stuff, man. Is it just that we're so nose down, like nose to the grindstone, trying to get work out the door that we're not hitting pause to invest in this Is it that you don't know where to turn to get good training? I can tell you as somebody more on the training side, it's very hard to develop training materials. I'll shoot a YouTube video with the plan of it, you know, going through our production process which takes about a week and a half and then eventually get published. Right now by the time it gets posted, there's already things that are out of date about the video. So doing a bigger lift, more deep dive sort of educational content. The idea of doing that is really hard because this all changes so quickly. So I'm trying to focus just more directly on like Day to day training content. It just doesn't work in like bigger production value YouTube things, because that stuff takes more time to develop. A great pair of quotes here. One from an owner, one from a staff person. Here's the owner quote. I don't see AI as a disruption. I see it as an opportunity. It can only be a disruption if it catches us unawares and we're unable to pivot. It's my role to ensure it does not. And then the staff person's perspective. AI can be very helpful in some areas. However, it's very concerning and scary to adapt to. And that's the voice of someone who hasn't been trained on it. That's such a big difference there. One, it's exciting, it's an opportunity. The other, I'm honestly a little spooked because it's pretty good and I can do some stuff that's part of my job. But again, the folks that are doing PhD level stuff with AIs right now, they are PhDs. We're scaling the output potential of smart people, folks that you already have. Does that mean it never replaces anybody? Probably not. But that fear is a product of not being plugged into what's possible and putting it to work. And that's just so, so important. We need to be doing that right now, putting this stuff to work, getting our teams up to speed. It's never been more important than right now. Have you got any ideas on how to, how to do that effectively with folks? Because it's extremely hard. Like I get asked to do a lot of AI talks and I have to turn a lot of that stuff down because I just can't do them all. And there is not a huge abundance of people that have both the domain expertise to understand how our work gets done, but also like real AI experts to be able to speak confidently and accurately about this stuff. A lot of folks are just parroting all sorts of other forms of thought leadership like AI being insecure and all these different things. So there's just not a huge number of those folks out there right now. I'll be honest, I don't know what a great version of this looks like because things are moving so quickly. But if you got ideas, man, drop that stuff in the comments. Because I get people that ask a lot. They're like, can you come and train my firm on how to use AI stuff? And I'm like, I can't really start down that path because that's not super scalable. And anytime I spend doing private stuff takes away from stuff like this where I can help everyone. And so this has always been my top priority. But when I look around, I also don't know how we solve this problem right now. The reality that people just don't have access to really, really good training materials. Kudos to the folks at Carbon for putting this report out. It's really good. It drives a lot of cool conversation, too. Just by tuning in today, you probably learned a few odds and ends. You got your own opinions. You may not agree with me on everything, but by talking about it, we're moving this conversation forward. Grab your own copy. Link will be down in the show. Notes, gang. Thanks for coming up. Being here today, it was great to see you. Looking smart. Looking smart, gang. Hey, I'll see you in the next one.