527 Does AI Mean ADVISORY Is The Future for Accounting Firms?

I don't know about you, I've never been more convinced than I am now that like AI is it is going to change a lot of accounting, a lot of the work that we do and that that feels unavoidable at this point. And it's leading a lot of people now to say you've got to start selling advisory work now because there's going to be a future timeline where that's all there is. Like when I We've already talked about ChatGPT's agent mode, going out and preparing a client's tax return in TurboTax and at matching the output of your professional tax software. And of course this is the worst it will ever be. So when we have, you know, desktop AI agents that can work, your tax software can go do stuff in QuickBooks. If we really get to the point where AI just kind of does all of that stuff, then is the only thing left advisory and if so, should it impact how we behave today? Also, because it's Q and A Friday, buddy, I rounded up some questions around AI hallucinations and how we should feel about that stuff and and a really big headache right now with AI tools. You're capturing your meeting transcripts, but how do you organize them into ChatGPT to be able to like usefully retrieve them? How do I share that access with my team? We've got this log jam of all these meeting transcripts. We don't know what to do, where to put them. We're gonna talk about that and what I actually think the future solution for that problem will be. Let's do it. Come on. If you're new here, welcome to the Jason on Firms podcast where we talk about running accounting firms, better ones specifically, and oftentimes that takes us to AI. I've got a bit of a software background. I've worked with a lot of brands in the space on their AI strategy. I help the AICPA develop their generative AI toolkit. And so I have seen some things and have some opinions on AI as it applies to accounting firms. As someone who has both run a 40 person accounting firm and done some stuff on the AI research side, here's a great question. Who's a podcast comment originally from Derek Datch? Maybe Derek said, hey Jason, do you think Accounting Advisory could get so high in demand that it could replace demand that previously belonged to automatable compliance work? It's a great question. Basically, if we get to the point where like compliance and basic stuff can be automated, then will clients just leave or is there really enough advisory work we could do that would backfill the work that goes away. Because if you think about like the entire market of services that accounting firms sell, I think most of us can agree a lot of the things that are happening right now, that will shrink, making the entire pie smaller. If you just look at that shrinkage in a vacuum. I won't giggle if you won't. But what Derek's asking here is advisory, could it really grow enough to counteract the shrinking demand for those more basic services? Here's my take, because people, there's definitely a growing number of people that talk about advisory as this imperative. And it's kind of has been a thing for maybe the last decade. I think it was super fueled by, in like the 2016-2019, we had this huge reporting tool bubble where a ton of money was going into reporting software platforms and those platforms were preaching the imperative that was you need to sell advisory work. And we promise you're just one good report from our software away from being able to sell advisory work. And as with most things, like there was a little bit of truth to that, but there was also a lot of money riding on the that hopefully being true. But if you look around, if you look at all the people listening to this podcast, are there people that sell advisory? Absolutely. But it's not the vast majority of accounting firms. And part of that is just behavioral, human nature. If we save time doing something, our default behavior will be to then just add more of the thing that we already do. Like the whole notion of automating tedious work so that we can do more valuable work. That usually doesn't happen. Like that's not the default. It certainly can happen if you're intentional about how that time gets reallocated. But the default for everybody is if I find a tool that helps me get more month end closes done faster, I will just do more month end closes. I won't take all of that time that I've saved and go and learn an entirely new skill and how to sell this work and how to communicate with clients in a completely different way that is not as natural as simply doing more of the same. And it's why advisory work, like as a category of work on its own, where people actually refer to it as advisory, it's maybe grown some, but not a huge amount in reality. A lot of people are doing advisory as part of just the day to day work that they do with their client, but they don't charge for it. They don't call it that. But what if that bookkeeping work, that tax work, what if it goes away? My take on this is I think it's very unlikely and that the way we do something now just completely falls off the map in such a rapid way that there isn't a simple evolution to like, well, here's what we're doing now. I think the progression of tech automating more of what we do also leads to a progression and an evolution of what quote, unquote bookkeeping is, and quote, unquote tax work and advisory work is what somebody would have called advisory 100 years ago is very different than what they would call it now. Maybe the better example is, is bookkeeping. And I've made this argument on the podcast before and I think I spooked some people. I would argue that bookkeeping already, by and large, has been automated. But that's not to say there isn't still a place for bookkeepers. I would argue that what bookkeepers are actually doing now is something closer to controllership. Because bookkeeping originally it was entering transactions. It was the logging of the records and classification. It was, you got a giant ledger book. Somebody has to put all of the transactions in there. But what do we have now? We have bank feeds that do all of that stuff. We have AI that's increasingly good at doing the classification for us. And even before that, we had bank rules where your highest output bookkeepers were leaning super hard into bank rules. So they weren't doing the vast majority of the classifications manually. That is what bookkeeping is. And it has by and large been automated. But have the bookkeepers been replaced by. No, I would argue we have more. We have a shortage of bookkeepers, of good bookkeepers. There's definitely a lot of people that come in and they're like, I'm going to fire up a QuickBooks file, and I'm a bookkeeper now. And they think they can just kind of figure that out. But I would say, like professional bookkeepers, we need more of them. And what most people would describe as a really good bookkeeper is actually doing something much more than bookkeeping. And whatever the label is, the role of an accountant, the role of an accounting firm, advisory work, bookkeeping. I think the most likely thing here is the definition of what each of those things do. It just continues to evolve as it has evolved in the past. The main difference is it's probably just evolving faster now. So will bookkeeping shrink? Will tax work shrink? I think certain parts of it in all likelihood will. But along the way, we sure seem to pick up a lot of other stuff. And a lot of people said this about the transition to the cloud. Well, all the systems can talk to each other and they'll all integrate perfectly and they can push all that data back and forth automatically. Because as sad as this is to say, a lot of what accountants do is taking data from one system, doing some logic and putting it into another system. Think about what tax prep is. There's so much of that, so much of keying stuff into your tax software. The reason why accountants insist on having multiple monitors is there's a lot of looking at this app and looking at that app and I know there's like there's a whole lot more to it than that. And it's the logic that happens in between that is very, very important. But back in, you know, as we were originally transitioning into the cloud and you first had APIs that made it so that apps could talk to each other in standardized ways, everybody's like, oh yeah, no, we'll have cloud tax software and all the data will get pushed in from the other system. You'll have your cloud accounting software and it'll talk to the bank and all the other systems and all the data will just go back and forth as needed and that'll remove a whole bunch of work. And in some ways maybe it did, in other ways it created more work. But. But yet here we are. Purchasing has gotten more complicated. The world's getting smaller, so more and more people are doing international stuff. Politicians keep making dumb compliance rules, making everything about being an entrepreneur so needlessly complex that you have to have an entire team of accountants supporting you just to be able to get your head around what is required as a byproduct of day to day operations. In fact, if you even just look at like income tax filing in the U.S. look what the tax system is today. This tax system could not have existed in its current form before computers like the complexity. They are willing to layer on scales with the tools and the tech that is available. And something tells me that will continue. Now like, to be clear, I am bullish on AI stuff and we talk about it a lot on the podcast here because I think it is an imperative for you and if you are a leader, your entire team, like you need to be pushing them. You gotta learn this stuff. Like you need to learn how to work with AI assistants and Chat GPT is the best place to start ChatGPT for business. That is the plan that you want. That doesn't come with the model training. It is imperative we get this into our team's hands so that we're learning the skill of how to work with AI assistants. But I don't see a moment where it's like the compliance work or the bookkeeping work just goes away. And then we're scrambling to backfill that gap with advisory, because I think all of it is just evolving. Bookkeeping is evolving. Tax work, compliance is evolving. Advisory is evolving. I think there's been a lot of examples of this in the past rpa. Like, there's been so many things that were coming for the extinction of accounting. And for whatever reason, we are more short of accountants than we've ever been before. And there's more work to do than ever before. The most important thing to me is like, how should this change my daily behavior? Nobody has a perfect crystal ball. And so looking out into the future all that much, honestly, doesn't really matter because I don't have control over that. What I have control over is what do I need to do every day, every week, every month? The people that are tuning into this podcast, I'm really not scared for you. We're talking about the really cool stuff out there. We're talking about ChatGPT agent mode and the cool things that it can do with spreadsheet automation. Talking about, you know, meeting recorders and the value of. Of capturing all of your meeting transcripts now, because you're never going to regret having those transcripts going forward. I'm really not worried about the people tuning into this podcast because in that next version of whatever running an accounting firm looks like, somebody has to pioneer what that is. And those people, like, they're. You look, look around in the industry. Where else are those people going to come from? They'll come, like, they'll come from other places. But the people listening to this podcast right now, like, you are already so far ahead of the curve on. On how you are thinking about this stuff. Even when I get down rabbit holes and say a lot of words that, that maybe you don't understand, Even when I get, like, way down rabbit holes that are just, like, way too technical, and you're like, reel back in, buddy. The folks tuning into this stuff, like, I think you are, you are a part, a big part of what the future of accounting firms look like in an age of AI, when clients have access to unbelievably good advisors. Like, if moving to advisory work really was the imperative that people would lead you to believe it is. Look at all the AI developments out there. Show me the AI that's helping normal people automate their month end clothes. Hard to find, right? But show me the AI that's a really, really good advisor. I mean ChatGPT is pretty darn good. Like it comes down to like, do you know how to work it? Are you giving it the right information? But I would argue if AI has come for anything so far for like a normal business owner, it's probably advisory, it's probably being a coach to you because ChatGPT isn't filing your tax returns, it's not doing your month end close. It can give you really valuable information to help you do those things. But at least right now it feels like while there's gonna be really exciting changes and all that stuff, the future is, is, is just an evolution of what we are doing now. And we have to learn these new skills to stay relevant and ride the wave. But I don't anticipate it being this like massive shakeup of the service offerings that a firm would, would put out there for clients. If you disagree, man, pop that stuff down in the comments. The, the, the most common folks that disagree with me on this are the ones that are just super, super bullish on all things AI and generally we point at other people's domains and say, no, it's totally going to replace what you do here. Just because we aren't plugged into the nuance of what other people do. It's why everybody thinks that the last person that's going to be replaced by AI is them, because they understand the nuance of what they do more than, than outsiders do. And this is why accounting for decades has been something where it's like outsiders kind of go, it's why is that still a thing? Accounting still a thing? Sure. Surely the next tech evolution will finally do away with that. Finally do away with, with putting together a tax return. Right. And there's just, I mean there's a lot of dead bodies on that road of software companies that thought they could come in and do that. There's a misconception that what accountants are doing is tedious and repeatable, like repetitious stuff. It's like, no, that's not what we do anymore. I mean, some accountants for sure do stuff in like old timey ways, but what we do is we do all the stuff in the cracks, all the messy stuff. We automate what we can automate, but then somebody still has to like get it to work together and fix it when it breaks. We're already at a place where we're not doing the tedium we're doing all the other stuff, filling in the gaps. Not to say we can't use more help from tech, that we can't get more automated in what we do, but the perception from the outside of what accounting is just the mental image that like accounting gives people. It's not really reflective of your day to day which is chasing clients for documents, fixing the stuff when it breaks, like doing all the nuanced stuff in the middle. Can I just air a grievance with you right now we are in peak tech selection season and your most important tool in your firm, it is your practice management system. But when you go to these providers and you ask them what firms are you going to great for? You know what, pretty much all of them say accounting firms, everyone. Our tool is the best for everyone. Buddy, that just ain't true. And one platform that is being built for a very specific type of firm, it is Client Hub. They know who they're great for and it is cast and outsource bookkeeping practices like accounting department in a box. Is that what you're doing for your clients? If so buddy, look at Client Hub. They're building for that very specific type of firm. Good example. They have like a, and arguably a deeper integration into the ledger than maybe any other practice management system because they give you like an entire way to manage the month end close inside of Client Hub. It's the sort of functionality that with most other practice management platforms you would have to pull in a separate month end close automation tool to get that functionality. But it's all happening inside of Client Hub because they know exactly what the firms who are a great fit for them need. Buddy, it is no different than what we talk about in accounting firms where it's way easier to serve a specific type of client at a very high level. That's what Client Hub is doing, man. They're building a platform for a very specific type of accounting firm. And if that is you, they really ought to be on your short list of practice management systems that you're considering rolling in to 2026. So to learn more about Client Hub and actually some really cool new AI workflow stuff they've got going as well. Check out the link down in the show notes let you know a secret about offshore hiring. This was, this was actually my experience and now that I've talked to more folks that have done it, a lot of people go through this progression. So in my account community realize about 40% of those firms, it's almost 600 firms, almost 40% of them are doing offshore hiring. And that is, that is way higher than that was like three years ago. World's just gotten smaller. I think Covid had a bit of play in that. Distributed work is kind of the norm. Right? So it doesn't matter where your team is. American accounting, it's not that much different than elsewhere. Even US techs, you can find experts around the world for that sort of stuff. But the team behind this video sponsor, Team up, they noticed the same thing that I did and that's that most firms go through a progression where they first hire a middle person to employ the staff and all that because they trust that service a little more and that service handles a lot of the logistics with the employee. But then you end up paying as much to that service as you do to the employee themselves. This is where most firms start with staffing. But once you get past like hire four or five, what I commonly see is they then just start hiring people direct. They're like, I actually just want to employ this person directly because half of what I pay I don't want to go to this other group. This is what Team up was born on. What Team up does differently is they will help you directly hire accountants in the Philippines. No middle people. All you actually pay to team up is a recruiting fee and that's it. So whether you're just getting started or whether you're maybe further along and you got now talented people that are trapped inside of these other organizations by the way, virtually impossible to get those people out of the those contracts. Team Up's a great way to find folks to come work for you directly. Learn more about Team up check out the link down in the show notes. AI hallucinations. All sorts of like comments and questions and all that stuff around AI hallucinations. I just want to have a quick conversation about this kind of summed up by one Internet comment quote. I will try AI when the hallucination rate gets to zero. This is a hard thing about working with AI is we're used to computers being perfect. We're used to being able to being able to trust the calculator to always give me an exactly correct result. But with AI it is not that clear. Like this is the greatest strength and the greatest weakness of generative AI is it can do more open ended things like I mean I can't talk to my calculator and have it do the sort of things that Chat GPT can do for me. But a side effect of that is it won't always get it right. In fact, what we call hallucination is the same thing as, like, its ability to be creative and come up with solutions. It's making things up. It's literally predicting each word or each part of a word what should come next, like one token at a time. And so the AI that can't hallucinate is the AI that like can't come up with creative solutions and think outside the box. And there's interesting information from OpenAI recently where they blamed some of the hallucination problem on like school testing, forcing people to guess when they don't have a right answer. Like if every multiple choice question that had ever been put in front of a student had had an option that said, I don't know, they say, OpenAI says our large language models would then say they don't know when they don't. Kind of feels like a software company, like just trying to dunk on the education system. But you can also kind of see that, like, where does it get this eagerness to always have an answer and always be a know it all? Well, it's like inherent in the way that we educate people. And when all that information gets pulled in the training data, the large language model's like, no, I have to always have an answer whether I actually know it or not. But there are so many analogies here to be drawn between an AI tool and a human that you work with. Where the same people, not the same people, but for the same reasons that some people struggle to make an intern productive in their accounting firm. People are also struggling to make an imperfect AI productive in their accounting firm. But I would argue the fact that we have quality assurance processes and like reviewers and all that, like accounting firms, more than a lot of businesses are actually pretty suited to plug AI in because AI is imperfect the same way that Carl is imperfect, the same way that you are imperfect. There's a study, I think I talked about this on the POD at one point, but studies maybe nine months old now, and the AI models they were using were like six months old by the time the study came out. But it was legal firms benchmarking those using AI versus not using AI. And what they found was the ones using AI were vastly more productive, but about the same in terms of accuracy. So oftentimes we think a computer will just, they will make us more accurate when that's not really what generative AI does. At least as of 12 to 18 months ago, it wasn't doing that. It has gotten much better. Hallucinations have gotten much better since then. But the biggest gain was in productivity, not in accuracy. And so the expectation coming in is, oh, this will just make me more accurate, or the AI is going to be more accurate than me. Like, that's not always the case. I mean, there's a hundred ways I can pick that apart where it's like, well, what if you just use it as a double check for what you're thinking and to pull information more quickly? Almost certainly you working with the AI like the end result is, is more accurate, but the reality is it won't be perfect every single time. Can we put it to use in valuable ways if it won't be perfect every single time? I think, personally, I think the answer to that is a resounding yes. For all the same reasons that we're able to work with imperfect humans. And so it's not about like, you know, the tools, the models, all that. They will continue to get better. Hallucin. Hallucination rates will continue to go down. But I don't have any control over that. All I have control over is how do I design the systems in my firm to accommodate for that, to ensure that we're still making sure that the output is correct. And how do I work with an AI in a way that helps me get to a provable answer. And so if you look at, say, research, I could cut the AI loose to do some research, and it just gives me an answer back. But if it didn't prove that answer to me and show me the sources and all of that, then I'm not really any closer to an answer. And so if you consider how would I cut an intern loose to do the same project and how would they prove the result to me? That's really what we want AI to do. How do, how do you give me like a provable answer back? Because the goal should not be get an answer as fast as I can. The goal should be get it through review as fast as I can. And it's why accountants have always had work papers and ways that you train team members to put things together in a provable way with references and all of that stuff, and all those same things you can do with an AI assistant. So it gives you a result that is more quickly provable. And so folks that are thinking through that, they're, they're working with AI in more efficient ways than the person that rocks up and probably uses the wrong model, that's, that is more likely to hallucinate. And they, they have a bad experience and they put it back on the shelf and they say, tell me when the hallucination rate gets to zero and I'll. I'll try it again. But along the way, they're not, they're not learning anything, right? The solution to me is just learning, like, it's just being vulnerable about maybe the things we're struggling with. It's why when I talk about the value of meeting, having a standing meeting with your team every single week to talk about how people are using AI, part of that meeting needs to be, show me your wins, but also show me your losses. Because some losses there are fixes for. In the past it's been. In the past, it's been people using the wrong model for a job. These days with ChatGPT, it's. That's actually gotten much easier because GPT5 is pretty good at doing just about anything. You don't have to pick from eight different models anymore. And so there were some situations in the past where you could give somebody a quick tip and it would then do it correctly. But there's other situations where it just won't work. And that's something for everybody to be aware of. Like, it just, it can't really do this very well right now. But unless you're having those conversations, nobody's learning that nuance because, man, nuance does still exist. Like working with spreadsheets, as an example, ChatGPT right now, if you enable agent mode, it's way, way better with spreadsheets than if you don't enable agent mode. And I wish it automatically enabled it every time you gave it a spreadsheet, but it doesn't. It's also way better at working with existing spreadsheets than it is creating a spreadsheet from scratch. But the vast majority of people, they ask it to produce a spreadsheet and they have a mental image of what that's probably going to look like. And the AI almost always does something different. And so you're not really happy with the result when the way better way to use it is to have a spreadsheet template for your recurring process. You give that to it and then it completes it for you. And then it's exactly what you'd expect because you already gave it the template. You can even secure the, the tabs on the sheet. So, like, only the fields that you want it to mess with can be messed with. But that sort of nuance, like, you don't learn that stuff until you, like, honestly have probably watched some of my content on, on YouTube around ChatGPT, but also have had those conversations with your peers week over week to just kind of keep leveling up and keep leveling up. And the really exciting thing is every time you find that cool new use case, it's like, sweet. I never have to go back to doing that manually again. And there's, there's so many examples of that. Like my favorite example we've done on YouTube with ChatGPT's agent mode is having this big old payroll journal allocation spreadsheet and you give it just the payroll journal export from the payroll software and then it puts it into this allocation spreadsheet over a thousand entries, does it flawlessly in like three or four minutes. And I see that and I think of all the different versions that we had of that for payroll and splitting, you know, credit card statements empty in different ways, we had so many different versions of that. And AI is now good enough to where nobody in my team would have ever had to do that manually again. Like we've crossed that threshold where there are things we do today that we will never have to do again because general AI tools are just sort of good enough and you can just give it some high level instructions and it will go out and do it perfectly. Like that's the power of investing and learning this stuff right now is you're finding those new things every week. I recently I had Chat GPT write a Google Apps script for me. I'd never done that before. I didn't even know how to pull up like a code editor. Basically what I had to do was export every Gmail email that I had ever sent and get a CSV of every single email address. And so that meant that I had to like combine duplicates and do all these different things. And ChatGPT gave me a like a 3 or 400 line bit of code. It showed me where to put it. I ran the code and in less than 10 minutes I had this CSV of like 5 or 6,000 email addresses. And there's no, I had no other way I could have done that. Like there's no mass, like bulk export of emails. Like I would have never been able to do that before without the help of AI. And now it's like, oh dang, what are all the things that are possible with Google Apps script that I wouldn't have known how to do before that. Now I can just get ChatGPT to do for me. There was a time not that long ago where all that was much more fiddly and the code usually didn't work on the first try. And now it's just like it nails it, it's so much better. That's, that's to me is the fun right now is like the more time you spend with this stuff, the more you're continuing to level up in a really exciting way. Have you heard the news from one of this episode's sponsors, Keeper. Introducing Keeper Scale. It is a whole new package of cool razzle dazzle stuff including automated accruals automatically tying out like what that schedule ought to be to the books. Rolling that forward month over month. AI journal entries hold up. It'll take complex source docs like payroll reports, Alta settle statements, Alta settlement statements, I don't know what that is and POS exports and turn them into review ready entries you can post straight to the ledger. Buddy, I like what you're cooking. Keeper smells nice. Keeper it is a month in automation platform. Keeper it is a month in automation platform for bookkeeping firms. It is like a task management had a baby with a ledger. Got a deep integration there. You can see all transactions inside of Keeper. Sort of one place to manage all of your month end click, closes. And they've been working on some pretty cool new stuff. If you're a bookkeeping practice worth learning more about Keeper, check out the link down in the show notes. Funny thing happened recently when I was at a conference with one of this episode's sponsors, Makers Hub. So Makers Hub, they're like an AI accounts payable platform. The party trick is they're using like AI vision models to do extraction at a level that was kind of never before possible with OCR technology. But I was doing something at their booth. Some of the customers came by like it is one of those tools where the customers are kind of real raving fans. But a funny thing kept happening. Like people kept throwing out these ridiculous document situations they have where they're like, I got this client, I have X, Y and Z problem. It's a total mess. We spend so much time on it and they just start riffing on it like with the Makers Hub team. And they're like, oh yeah, no, I can see like you do this potentially. Like our product doesn't really do that. But actually if we change this or that this could potentially work. The team is just like, they're just a bunch of like document processing who are also AI nerds. And the product very much is like the marriage of those two things. And it's what Maker Sub does best right now. Like the biggest, ugliest accounts payable stuff that you got, man. I had some gnarly clients where Makers Hub would have Saved us so much time. If you're out doing conferences at all, you'll see Maker's Hub a lot. Add Intuit, Connect. I'm gonna be there. I think. I'm actually co hosting a thing with Maker's Hub. But if you got any ugly bill pay stuff, class tracking projects, inventory manufacturing, stuff like that, definitely worth giving Makers Hub a look. To learn more, check out the link down in the show notes. Okay, last question. We've gotten a few different forms of this. Basically, you're using an AI meeting recorder. Awesome. Means you're getting transcripts of all of your client meetings, maybe even your internal training meetings. Those are super helpful when you got to go back and be like, did they mention this or mention that? And you can chuck it into an AI tool and ask it, did they talk about the rental property? Did they talk about who's claiming the deduction for their. The credit for their kid this year or whatever that is? That's a vastly better version of finding information inside of a bit of text than like the old control F. Find a specific word, right? And you can imagine a world where you had all your transcripts for the last 10 years and just how valuable that would be to be able to query that stuff and like, all sorts of things you can do with that. So AI note takers, very, very popular now, a lot of firms using it. But where do you put those transcripts in, you know, say, chatgpt in an organized way? Do I put all of the meetings for a specific client into ChatGPT? Do I do that as a custom GPT? Do I do that as A project? One update here. ChatGPT has projects. Historically, projects could not be shared. Now they can. And so a project is kind of just like a folder and you can upload any amount of information to a project, and then whenever you chat within that project, it can see into all that information. That's certainly one way that you could organize client transcripts, is have a project for every client and then share that project with the rest of your team. The rub is, I mean, I had 1900 clients in my firm. I don't have 1900 projects in my chat GPT that are shared between team members. That's a mess, man. And so the honest truth here is they're still in. ChatGPT is not a great way to organize this stuff. We need to be saving these transcripts to the client file, to the project perm file somewhere internally. But inside of Chat GPT, there's not a very helpful way to store this stuff and then share it with your colleagues. And so what we end up doing is more sort of ad hoc. I need to look into one specific thing and so I'll create a project for that specific thing. That's probably still what I would recommend. If you had a need, let's say you're having turnover in your team and the accountant who's going to start meeting with that client every single month. They're new to the account, but you've got the transcript for the last 12 months of meetings that the previous person ran through with that client. That's really, really valuable. That's a, a great way to train that person and get them up to speed quickly is to have all those past transcripts. If you had that come up, I think then you go to chat gbt, you put all those transcripts into a project and then you share that project with the new person. And so like right now, organizationally I think you're still kind of doing this stuff. One off, like there isn't a perfect answer for how to organize all of these things. I hope this will change. There are two people working on this, two note takers specifically for accounting firms. AI note takers. One is called Vinyl, the other is called Ping. I'll put links in the show notes to both of them. First they are like rolling out integrations with your practice management system, which is super nice, the notion that you can have a meeting and it'll automatically push that transcript into your practice management system. Awesome. But then the next step is more around like how to do intelligent, useful things with those transcripts. And so a basic feature of these tools is you have all of your clients in the note taker. It knows who you're meeting with and like what clients those contacts roll up to so that then these platforms can be more oriented around like giving you feedback on meetings and actually being the home to where you chat with these transcripts down the road. Removing the need to pull a bunch of transcripts into ChatGPT and devise this whole complicated like storage structure in ChatGPT. As much as I love ChatGPT and, and I do think you need a license ChatGPT for business in the hands of every member of your team. Because it is still the ultimate Swiss army knife that can just do so many different things, I don't think long term it will be the place where you talk with your meeting transcripts. I think that's either going to be more purpose built meeting assistants like Vinyl or Ping like the ones that are being built for accounting firms or eventually this could just end up happening in your practice management system. If I push the meeting transcript to my practice management system and I've got a good like AI chat assistant inside of the practice management system that can see into all that stuff, I could definitely see that being where we're having those conversations as well. But right now like in my opinion there is not a great way to organize that stuff in ChatGPT besides just having one off conversations when you need them. Wish I had a better answer than that, but also glad to see that we have folks building note takers for accounting firms too and like building out the integrations with the rest of our tool. Very exciting. Again, links to those two down in the show notes Gang, thanks for being here today. You got any questions? Man, pop that stuff down in the comments. When you ask a question, somebody else has it too. We talk about it, we share some ideas and like we all kind of are thinking further down the path after flushing out some of those thoughts. So keep those questions coming down in the comments and I'll catch you in the next one. Foreign. Hey, real quick for my podcast friends. We just announced 21 live events were running between now and the end of the year, starting with dinners in Denver, Chicago and New York City this week. Get the whole team out if you do tax work, celebrate getting past the deadline. Right from there we're headed into a Connect to Vegas. After that, events in Toronto, Orlando, Austin, Oakland, Oklahoma City, Phoenix, Seattle, Portland, San Francisco, Louisiana, Atlanta, Nashville, Chicago, Washington, D.C. philadelphia, we're back in New York City again and then Boston. And 13 of these events, they are what we're calling Jason on Firms Live, a half day workshop experience that rolls into a dinner that I am so excited about. We've built these ones from scratch. These are events where I want you to bring your entire team to get everybody rowing in the same direction headed into 2026. If you're an accounting firm leader that has ever struggled with change management, with having all this enthusiasm but not being able to like get the team to commit to your vision, it's the entire reason we built these events. If I've ever inspired you on this podcast, the idea with these events is that we can now inspire your entire team to get everybody working in the same direction so you're not putting the whole thing on your own back. We'd love to see you out in the wild. Head out to jasononfirms.com events. You'll find up to date information about everything that's coming up. I'd love to see you there.