586 First Impressions: Claude Cowork for Accounting Firms [Is this the most powerful AI agent for accountants?]

Today on the podcast, we're going deeper on the AI agent that many people say right now is tanking the US Stock market. Definitely feels like we've hit a turning point. Like AI agents right now are becoming more real. But where are they at for accounting firms? Like, are they to the point where it's something that every single member of my team should be using? Today we're going through some of the recent stories, but also looking at a really deep dive that I've done in the past week, specifically on Cloud Cowork. As we're putting together a big video on sort of the do's and don'ts of how to use that AI agent. I want to show what's working right now, what isn't, and what this tool actually means for accounting firms. It's very rare that I say this. Right now. Most people are underestimating the impact it will have on how we work in accounting firms. Very interesting. Also today we're going to look at a really interesting niche that is like just the sort of thing I would build my accounting firm around, but specifically what makes it really good. And my goal is you'll come away with some fresh ideas regardless of who you are going after in your accounting firm, of how to find an even better client and maybe just have a better framework for knowing, like what even is a good client for my accounting firm. Last, we're going to rummage through the mailbag. Got some great questions around. What are the best a AI agents for accounting firms today? Some of the dos and don'ts of chat GPT's agent, how to start an accounting firm, if you're thinking about it, and a very specific relatable example of, of something that firms struggle with to get clients saying yes to better prices. That is, that is a bigger number, a better, a better price for you. All that on today's installment of the Jason on Firms podcast. If you're new here, welcome in. What we do here is we talk a couple times a week about running better accounting firms. Whether you run a firm, whether you work in a firm, or whether you one day aspire to know the pain that is owning your own firm. The pain. There's, there's definitely painful days as there is in any sort of entrepreneurship. Okay, Claude, Cowork. It's not new, I'll say that first, but I will say it is newly entering the awareness of more of the world. In fact, increasingly I'm seeing news stories that like tie this sort of next wave of AI agents to stock market issues and I can kind of see it like the timing does line up and more than ever before, there is just a surging of confidence of surety that knowledge work is about to really, really, really change. And I'll be totally honest with you, and I've been in big time AI dork from the beginning. I mean, the beginnings of this podcast were shortly after ChatGPT launched. Been talking about AI stuff from the beginning. I have a bit of a software background myself. Until two weeks ago, I wasn't taking this stuff seriously enough. And what's changed is we're in the middle of really deep dives on Cloud Cowork. I feel like we're pretty much through the cloud cowork deep dive right now as I speak. We're doing a deep dive on, on OpenCLA, specifically around accounting firm applications to be. None of these agents right now are ready for client work, are ready for handling client sensitive stuff. But I underestimated just how good and just how useful both of these AI agents already are to kind of a shocking degree. Today we're going to talk through a bit more background on what Cloud Cowork is. If you're not familiar with it yet, but also in our testing and the benchmarking that we've done, just the sort of things we're able to get it to do now that are kind of blowing my mind. When ChatGPT first came out, I lost sleep, like, because that thing blew my mind so much. I still remember, just like every new conversation with ChatGPT was like, be my accounting system now, be my invoicing tool now. Like, I would just keep asking it to do these things and it would seemingly do them and it wasn't perfect at the time. And I would say Sam Altman, CEO of OpenAI, he says the same thing. When you get a new model or something, some big upgrade, it's never as smart long term as it initially seems. And when ChatGPT first came out, there's definitely a lot of that. There was such confidence. You're like, what am I even talking to right now? This is incredible. And that has a way of sort of tricking you into thinking that it's being more intelligent than it actually is. Again, also not unlike other humans can do, right? Especially if you hire people. You ever sat, sat through a, a really strong interview and that person comes to work with you and you're like, holy geez. That was a, that was a paper thin facade. I lost sleep around ChatGPT because I didn't know what the boundaries were. On exactly what that could and couldn't do. I could tell you the last couple weeks I've lost sleep and I've never been scared. I would never go as far as to say I'm scared for accountants and what accountants to do because the overwhelming shortage right now, even in a really strong AI timeline, it never felt like you'd really end up in a spot where it was like, well, what are accountants going to do? Because there's just so much to be done right now. I've changed on that. I wouldn't say that I'm scared. I will say that I'm a little nervous about our abilities to change in the ways that we will need to change in the next couple of years. And I have never said this before, and I'm always very careful to sugarcoat it because everybody want, like, I do a lot of speaking, what everybody wants in their speaker is for somebody to come and tell them, everybody there that everything will be all right. And I never do that. And I think we have to be careful to shy away from the statements that feel that way of like, oh, you're never going to be replaced. You're going to be replaced by somebody. You know, accountants are going to be replaced. You'll be replaced by an accountant using AI. I'm really allergic to things that feel good, things that roll off the tongue, have a way of tricking you into thinking everything will be all right. And what that actually leads to is inaction. It tricks you into thinking I don't need to change my behaviors, my habits, and how I operate right now. When, when today. I'm a little concerned at just how much I think is going to have to change in the next 12 to 24 months. I'm not like a fear mongery guy when it comes to talking about this stuff. But I can tell you right now, it's been a time since I've felt such an obligation to get back to research, to get back to, like, practical use cases and actually doing stuff. Because in the last few months, what's possible with AI has completely changed in a way that practically nobody's caught up to yet. So to get practical here, to get down to like, exactly what is changing, what does this mean for us as a storm rages outside? Man, if you're in the northeast, it has been a wild weekend. I mentioned I'm not sleeping. 10% of that is AI stuff. 90% of that stuff is being sick. You got three kids this time of year, man. Your house is just a. Just a petri Dish. Okay, what's changing? Number one, what Claude Cowork unlocked is rather than just like the Claude AI assistant, which has been around for a really long time, Claude not too different to ChatGPT, run by a different company, this online chat assistant generally been very capable, but not quite as good as ChatGPT. In the last three months they've had a few, a few now new models that are tremendously good. But what cloud Cowork is, is a different harness is what they would call it. And so if a, if a chat assistant is, and this is a simplification, if a chat assistant historically has been you talking with a large language model, Cowork is the next step in complexity in the harness of what you're talking about. So you're having a conversation with it and what it's actually doing is on the back end like spinning off all these sub agents and sub processes and they can now even run in parallel. So as you're working with Cloud Cowork it can kick off a whole bunch of things working in tandem. And then what you are interacting with is almost like the orchestrator of all those things that are going on behind the scenes. If you couple this with better AI models and harnesses that will allow, that will now allow for you to do longer time horizon tasks. That is, it will work on a big project for longer, then it represents like a big change in the use cases that are possible because so often accountants are like, yeah, I threw this big project at it and it was useless AI dumb. And the first thing I always say is like get to understand with a chat assistant, it is hardwired into those assistants how long they're allowed to work on something. Because working on a problem means burning resources, means making server farms go vroom. And we do not all have an unlimited power to be able to do that. Like I can't for, for good reason. ChatGPT won't let me give it a 100 hour task. It is allowed to work and put in a certain amount of effort on the task that you give it. And so if you give it something that is bigger than it's allotted amount of resources, then it will either do it 20% of the way and just stop or it will like do this really haphazard sort of broad brushstroke version of it. We've, we've probably all experienced a version of that, right, where we give it a job and it comes back and you're like, like none of this is done. This has now gotten a lot better as we're able to let projects run for longer. So first, big change, specifically with coworkers, you now have like a novel new harness that will manage several sub agents behind the scenes and allow for things to run for a longer period of time. Second, big change is those processes can be built into what are called skills. So you show it how to do something and you have to like, kind of painstakingly describe xyz. And that's a little bit annoying. And that's kind of one of the big downsides of traditional AI assistants is every conversation feels ephemeral, like it feels like you're talking to a goldfish. Unless you're using a custom GPT or project or something like that, you're kind of starting from scratch in many ways, right? Accounting. There's so many things that we do for the same client, maybe month over month, or the same thing that we do for a bunch of different clients to where to go back. And it is like explaining to an intern for the first time how to do something, but starting at that point, every single time, Skills let you sort of package whatever that process was. And it's so trivially easy in Cloud Cowork to build a skill. You literally go through a workflow and have it do a thing and you just say, now create a skill to be able to do this in future conversations. And Cloud Cowork has a skill that is called, you know, skill Creator or something like that, like a skill for building skills. And then that skill, skill, like where you ended up in that project, can be used in the future in other conversations. And so if it's something that you do with a lot of different clients or something that you do for a client month over month, it makes it super turnkey. And so the next time you want to go and do the same process, it's vastly more efficient, more intelligent, and you can even tell it after that time, like, hey, modify the skill to be more efficient or incorporate any changes that maybe should have been part of that skill. Now that you've had sort of your second run through this process now those skills can be packaged up into what are called plugins. And plugins can be shared between you and I. They can be shared anywhere. There's a bunch of plugin packs online that the accounting stuff is very weak. In fact, Claude has a sort of finance and accounting plugin that was very obviously not made by an accountant. I think it stinks. But to give you one example of just how powerful skills can be is one of the use cases that we'll have in our cloud cowork video that, that I'm hoping will go live next week. A really big project. I've got this, basically, this spreadsheet. It's like this basic accounting ledger, and we all probably have some version of this. It's this Excel template where basically you just put the transactions in and then you put a letter into one of the columns and it codes it to, like, an income or an expense account that's in one of the other columns, basically just to do, like, serve as, like, a really quick and dirty accounting ledger for a client where you've got some transactions, but you can't justify, you know, spinning up a QuickBooks file or something like that. We've used Chat GPT's agent mode to complete this accounting ledger in the past. And ChatGPT's agent mode is actually very good. Like, it will handle upwards of a thousand transactions that it will enter into that spreadsheet and encode it in a way that's like, okay, like, maybe as good as an intern. Like, not amazing in terms of the classifications that it picks. And so we tried the same thing with Cloud cowork. Give it this spreadsheet template, basically, and then give it some bank statements and say, enter the transactions from the bank statements into this spreadsheet. We first did it with 67 transactions because I'm cringe. It did that without breaking a set, then without breaking a sweat. Then we did it with 670 transactions. And it took 12 minutes for it to do that. But I wanted to try it with 6700 transactions. Now 670 took 12 minutes. So 6700 transactions going to take forever, right? Instead of just chucking a huge volume of transactions into it, what I did after the 670 transaction test is I had it create a skill to describe to itself how to most efficiently enter transactions into that spreadsheet. And as it turned out, the what was taking the most time for it to enter into that spreadsheet was stopping on every single transaction and being like, okay, this is Home Depot. Here are all my classification options. What should I categorize Home Depot as? And then the next one is like, you know, some Etsy seller, like, oh, here's this Etsy seller. I gotta go research this. I'm actually going to run a web search and try to find this Etsy store and see what sort of stuff they do, then come back, try to figure out how to classify it. It was doing that on every single transaction, and that was what was making it take a Lot of time. I mean, still only 12 minutes for 670 transactions. Not that bad. But what we worked out in the process of like, hey, let's build a skill to make this more efficient, was it said, I basically just need to build like this sort of vendor directory because you actually had like 25 Home Depot transactions. And I went out and looked it up every single time. Let's build this directory so that anytime I see a new vendor, I go out and look it up. But if it's a repeat vendor, I don't just basically a bank rule. And so it now, like, it created a skill so that it would then sort of figure that stuff out on the fly and run more efficiently. So with that skill, I reran the 670 transaction test and it went from 12 minutes to three and a half minutes. I then put 6700 transactions into it. This was 12 months of bank statements that had 6700 transactions in those 12 months of bank statements. I can't remember how many pages of PDFs. It was a lot. I put those 12 months of PDFs into it. I gave it the Excel template and it completed the bookkeeping in 12 minutes. Using that skill that we created. 6,700 transactions across 12 statements. Put all that stuff into Excel. The classification, honestly, was like flawless by my eye. Even delineating between like office expense and supplies and kind of some of those fringe things. And it gave you some feedback around like, hey, just so you know, I did this with this and that and sort of these fringe scenarios. And it's like, that's a very intelligent thing, like just in the way that it described. Just so you know, the FYIs were so intelligent that it gave me a higher level of trust in how classified things. That blew me away. Not only that, that would have been impossible without the skill. And so that's, that's the importance of skills now as a way of making AI agents more efficient. But that it was able to handle that volume in 12 minutes. It was only last week that someone was on Twitter saying, like, how do I get chatgpt to extract the transactions from this bank statement? I just can't get it to do it right. And like, that is AI right now, man. Like, there, there is so much learning to it and a hundred wrong ways to do something in one way that will work. And the people who have spent the most time with the tools know the one way to get it to work. And then the other people come in, honestly, usually on the free tools. In fact, you May have seen this image going around lately. It's a question of like, I have to wash my car, like, what's the best way to get to the car wash? Or something like that. And AI gives kind of a dumb answer in a roundabout way. And everyone's like, ah, AI is so stupid. This is the sort thing that, that humans will always understand that AIs won't. And in that screenshot, you can see a login button for Chat GPT, which means the thing we're making fun of is the free version of ChatGPT. And all of the free versions of every AI tool just totally suck. So you'll see, you'll hear me say something like this. Like, I could do the bookkeeping for 6,700 transactions and pull them from the bank statements and put them into a spreadsheet and glass correctly. And then people are like, okay, I get it, I got to try AI. And then they go and like check their first prompt into this free tool. They're like, this thing sucks. And they think the AI thing is kind of a hype and everyone's, you know, blowing it out of proportion and sort of the cycle continues. Got some really, really cool use cases, really cool examples in this video. Similar tax examples where we ended up at like, I think 65 pages of PDF tax documents that it could enter into a 60 tab Excel workbook flawlessly. In fact, I even went back and it ended up being like over 350 entries into the workbook from the tax documents. I even went back and I changed 10 very, very small things in that workbook. Things like the state payer ID. I changed a single digit. One of the 1099 divs I would, I would remove. I added foreign tax paid to one of the 1099 divs. I changed the Social Security payments by a dollar. There were big old brokerage statements in this thing. I changed a variety of things from small, fiddly, tiny things that maybe don't even matter to like completely omitting a thing or adding a 1099 it from Yamama LLC. I made 10 changes. I put that entire workbook into cloud cowork, like a new conversation that didn't have the context of the past. I gave it that workbook along with 65 pages of PDFs of tax docs, and it identified all 10 things, just those 10 things. It specifically called out the 10 things that were itty bitty that I changed. And so your head kind of starts to spin. Like, I mean, I'm like, I'm a 15 year tax pro, I spent the last 5 years in my firm building a cast practice to a couple million a year. All I've ever known is doing accounting firm work and hiring talented accountants. Like there's no world where everybody isn't using this or something like this heavily every day by the end of this year. Now the rub with cloud Cowork today is it's still quote in Research Preview which means it doesn't come with some of the same data protections as the cloud team plan which is just it's chat assistant for business which is very secure and I could, I would happily trust with client sensitive information cowork right now still in Research Preview they say don't use it for regulated workflows. In the US Tax work is definitely a regulated workflow. So it still feels too early. The hard thing is a lot of the risk with these right now is just in browser use is in sending an AI agent out to the open web web because it can be tricked by things not too dissimilar to like phishing in an email. How a member of your team could be tricked by phishing and by the way AI agent could also be tricked by phishing. You could get an email, it could try to process it and do something it shouldn't. But a lot of the risk with AI agents and we're actually next week we're going to go deeper on like exactly what are the riskiest things about AI agents because understanding, you may have heard prompt injection, that's kind of the big phrase that's going around. There are other risks and understanding them more and how to mitigate them is going to be so key for the future. And making the most of AI tools so clockwork, very, very exciting. And that's kind of the, that's kind of the whole point of the video as well as the open Claw video that we're working on as well as I want to show you just how powerful it already is. But also understand today if it's me, I still can't put client sensitive information into it. But what we're now waiting into is such a messy transition of the humans being responsible for 100% of it to the AI being responsible for 1% and then 2% and 3%. Like just to know what's the right 1%, what's the wrong 1%. Well at least right now it looks like the work on the web, that's the hardest stuff to let it do autonomously. But I've got, if I've got an Excel file And a pile of PDFs that it can power through and do all that stuff. Awesome, Amazing. Like, to be clear, the model is still running in the cloud. So like all the, the, the knowledge and all the smart things that it's doing, it's all happening via an LLM that is in the cloud. And so that information, it's going back and forth to the API, just like it is if you're using, you know, ChatGPT. ChatGPT is a user interface. Behind ChatGPT. There's an API where every time you submit a message, it goes to the API and then the API gives a response back and it shows it in the user interface Cloud Cowork is no different. It runs locally on your local machine and it can see your local files if you give it access to them. But the intelligence is not local. Like that is all going to the cloud. But right now I don't really have a problem with that. We have completely secure ways to use ChatGPT. For example, you doing tax work right now, you got a lot of 1040s you're trying to get out the door, buddy. Have I got the app for you. Is this episode sponsor Sorban? If you do 150 or more 1040s a year, you need an AI 1040 intake automation system solution. Vastly better than Cognito Forms or the Paper organizer or anything that a practice management system has on offer. Right now with Sorban, you just dump all the docs that your clients are giving you into Sorban. It goes through a checklist of all the things that you should have for that 1040 based on pro forma data, automatically checks off all the items that are in those documents and gives you like an intelligent portal you can feel good about putting in front of your clients. That will save you a tremendous amount of time. And organizing all the information that you're getting in and figuring out what you have and what you still need and following up with clients, automates, all of that stuff. A tremendous tax season, time savings. We are in the thick of it. And this is where Sorband is going to save you a tremendous amount of time. I know tens, if not 100 plus firms now that are very, very happy users. To learn more about Sorban, check out the link down in the show notes. Can we talk about a very real problem with most offshore staffing solutions? Actually faced this one myself is the staffing groups that serve as sort of like a middleman, a middles person between you and your staff. Like you pay the middleman and then they pay part of that to the staff person. Did you know in most setups there's actually no way to get that staff person out of that agreement. If you're like, I'm actually tired of paying half the monthly cost to the middleman, like half the money I'm paying is going to this company, not going to the person. Can I just work with that person directly? Most of those contracts, there's actually no way to get that person person out of that employment agreement. That is why this episode sponsor Team up exists. They're an offshore staffing solution where you just, you work directly with that person. Team up, they're basically just going to serve as the recruiter and then after that there's no ongoing fees. Half of the money's not going to somebody else. You're just working directly with a talented human being. In the Philippines, the progression honestly for most accounting firms is they'll hire, you know, three to five folks offshore, have good luck and they're like, okay, what does it look like to go to 10 or 20? And they're like, I don't want half the money I'm paying to go to this other group anymore. I just want to pay these people directly. That's what Team up will help you do. Finding somebody great for a one time flat recruiting fee. Actually founded by a former bookkeeping owner too, if that's something you want to explore. Learn more about Team up down in the link in the show notes. And so there are like secure ways to work with AI. The bigger problem is with the power of something like Cowork and having access to the web, the ability to kind of go out and do things if you give it permission to. Like, the hard thing right now is rolling that out in the context, context of a business where it has all these features and you might know what features are okay to use and what ones aren't, but your team does not. Also, your team has very different incentives to you. If they got to get to little Tina's cello recital by 3pm and it's noon right now, and they just came through grad school and used AI the entire time and they know how to be productive with AI and they're like, well, I got, if I just had the AI agent go and do this, I'm, I know I can be out of here until the recital by three, so I think I'm just going to do that. Like, like their incentives like are not always going to be the same as yours. And so even if Cloud cowork is 90% fine for my team to use in any way that 10%. If it's sitting there in the app, which it is like they can cut the agent loose to go and do anything on the web, then how do I draw the boundaries around that? Like, we really have that problem with ChatGPT right now. I would say ChatGPT, the business plan historically, for the last two years since it first came out as the Chat GPT for Teams plan, has been secure enough to use with sensitive client information. And then six months ago, they introduced agent mode, which can go out and do stuff on the web. I can't let ChatGPT go out and do stuff on the web because it knows like, memories from other transact other conversations, like prompt injection. It. It is a real thing that we need to be mindful of. But if that feature goes live in ChatGPT, when I've rolled out ChatGPT to 20 of my team members, you kind of put me in a tough spot then, right when I have to be like, hey, just so everybody knows, like, agent mode, here's what it's okay for and here's what it's not okay for. And we are struggling with this with Cowork right now. It's incredibly powerful, but not something that I can feel good about rolling out across my team. Open claws like that times 10, open claw. The big difference there is it has access to like, everything on the machine. Like, it can. It's. It's like an incredible amount of control to the degree that, like, when we're doing this open Claw video, we're literally putting it on a clean install, like, operating system with nothing else on it. Because if I, you know, for example, if I plug it into my email, it can do anything and all it takes is it getting confused. And it's actually. You may have seen the news story of Meta, you know, kind of the. The parent company of Facebook, One of their AI alignment researchers, connected Open Claw, this new AI agent, to their email, and it like, deleted all of their old emails as they were telling it not to. Like, it got confused and went through and did this whole thing. And then the lady goes out and posts about it on Twitter and everybody's like, hang on, aren't you the one that's supposed to be saving us from the AIs? And you're just trusting it to manage your email and it like kind of blew up in your face, right, that stuff. Because it can happen. Right now. We just got to be incredibly careful about how and where we use this Stuff. And it's that conversation that I think is incredibly important because anybody can go out online to social media and be like, you can't use AI. AI is not ready yet. The better version of that conversation is here's where it's at right now and here's where it looks like the boundaries are developing. Because, gang, this is like the next 10 years of being productive on a computer is understanding this stuff and like saying like, can't use it, not ready, like, that's not learning, that's not getting any closer to a solution. So I think we have to be getting our heads around how these things work. Because what I just described, from the bookkeeping example to the tax prep example, like what I know about Cloud Cowork right now, when I employed, you know, we had a team of 40 and about 30 of those people were doing tax work. I would, had I had Claude cowork at that point in time and we could trust her from a security standpoint, Cowork would have been my most productive tax preparer. Genuinely in its current form, we would have used the heck out of it. Like that is stuff that we spent so many human hours on doing. Pulling information off of tax documents, putting it in the work paper templates, like it's already there. The problem right now is getting the security to a point where like we can clearly set the boundaries of the right ways to use it within our firm. But that changes firms big time. And right now, like it doesn't look like it's slowing down. It's sort of kicked off another arms race where OpenAI is now releasing models faster to kind of try to catch up with what anthropic is doing. OpenAI just hired the guy that made Open Claw and that I'm sure was an astronomical amount of money they just spent to get that guy to join OpenAI. And so Open Claw, this like kind of weapons free AI agent that could do anything on your computer. Like the folks at Chat CPT are like, this is part of our future. I can imagine. Imagine a chatbot type agent that can see and do anything on your machine from moving the mouse to typing on the keyboard, like literally anything. Imagine OpenAI releasing something like that. Like that is in the near future. And so the day it drops, what are you going to do? Because it's one thing to get the thing, it's another thing for the thing to be appropriate to use in my business. It's another thing to get it to where I can put it into the hands of my entire team and help make them more productive. And an interesting wrinkle here is you won't hear your software companies talk about this. And this is a this big deal, because 90% of the thought leadership content around accounting firms is put out by software companies. If you think about it like, and it's not bad stuff, it is just inherently biased. And what software companies don't like is this is the greatest existential risk to what they do that there has ever been. And so if your watering holes are, man, are a combination of a bunch of software companies that put out thought leadership, a lot of which is really, really good, plus human beings who are incentivized by software companies to put out thought leadership content. And I can tell you that is a lot of my LinkedIn feed right now of messaging being sent out through accountants from brands. That messaging, it's not going to say, hey, there's some really powerful AI agents that are perfectly capable of using this software. Like, you know, these apps that you're buying, like, it can actually go in and use the software. And so the day that an AI agent is just as good as me at using QuickBooks or using my practice management system or anything else, then that software just became a slow API. Like APIs are how software programs talk with each other. It's kind of like the interface you have for Bill.com to talk to QuickBooks to push a bill into QuickBooks. But if I'm talking with an AI agent and that thing goes out and researches some tax thing for me on Bluejay, then if it has to use the app, then all that app really is now is a slow API because I'm not using the app. The AI agent is. And so this, we've talked about this before on the podcast, this new sort of like AI agent layer of software where you have this agent that's above all the tools that you use. Very, very useful for you. And I completely changes the model of how software companies make money. We are truly moving to a potential timeline where you either are the agent or you serve the agent. And so if these software platforms, let's say QuickBooks puts up bot protection so that AI agents can't use QuickBooks, is that the right strategy right now or not? Well, if they do, what's to stop some other platform from coming out and being like, hey, work QuickBooks for agents. And now anytime somebody asks an agent to do something, it goes and it uses this other platform which speaks and integrates and works really nicely with AI agents. Isn't that kind of the future Isn't that kind of where we're headed? That's kind of like the Catch 22 that software companies are stuck between right now is we're nearing the point where the tools we use will have more, but more AI agent users than human users. And this is true for, I mean, pick a percentage, not 100% of software, but maybe 80% of software. And so if the way that we learn and the watering holes we hang out in are ones that are, like, full of thought leadership content from our software companies, like, this is not gonna be part of the conversation. In fact, what you're gonna see a lot of is people pooh, poohing the AI stuff and being like, well, like, being an accountant is about trust and it's about the relationship. And all of that is true. And if you take out all the work, and we talk about this on the podcast all the time, is if I'm running an accounting firm and can't be bothered to respond to a client's email because I'm too busy doing their work. I've been in that place, we've all been in that place, right, where you just can't be inconvenienced by the client relationship management because there's so much do and don't they know that they're just keeping you from doing your best work? The problem is the, the client, the customer, like, they're only making buying decisions based on trust, based on the relationship. They're not making it based on the technical stuff that they don't understand. And so when the AI can help us do all the technical stuff much more efficiently, the same problem remains. Can I take great care of these people and help them try to understand the complexity and the value of what we're doing for them so that, that whole, that whole bag, like, is still there? Even if AI is able to automate a lot of this work, if there's anything that gives me fear, it's the speed at which, like, how we get our work done is probably set to change. I'm really excited to see your reaction to our Claude Cowork video. I know I sound like I'm on one today, and the one that I'm on is dayquil on some other stuff for sure. But I'm super excited for you to see these videos that we've got in the can. I'm not usually hyperbolic about AI stuff, and I'm. Most people see me as like the AI accounting guy. It's just taking a big step. And I honestly, I don't Think we're talking about it enough. There's a tremendous amount of noise online and everybody's just talking about it for the sake of talking about it, to be part of the conversation, not a lot of people moving the conversation forward. Which to be clear, it's hard to do. Like I sunk probably an entire working week into the research and the production, all that that went into our cloud cowork video and now it'll be in editing for probably two or three weeks. It'll be a 30 to 40 minute video. It's hard to get your head around this stuff. It's a lot of work. But like I shared earlier, this is the first time in a long time where I've felt almost the moral obligation to like go heads down into research to help people to see the importance of what's happening and start driving more conversation about what are the appropriate boundaries. It wasn't that long ago that there was no secure way to use an AI chat assistant. And everybody just said AI is not scary, is not secure. And we really pushed that conversation of hey, let's get more, little more first principles about this. Because the fact that there are bad ways to use AI does not mean they're not completely secure ways to use AI. And if you look around now, that's actually like common wisdom is like, oh, as long as I'm getting the business plan of this, this AI tool, like I'm probably fine. And this is something we've been saying for probably two or three years now is like going deeper into like, but why? Like, what are the things about it that aren't secure and are there any ways to mitigate those things? We are heading into a whole new era of this with AI agents that I would say have much bigger stakes than the AI assistant conversation had because the capability of those tools is much, much less than being able to spin up multiple like cloud cowork things all happening at the same time. And working on these bigger, I mean, preparing it to extrajourn, like just working on like bigger lift sort of work and it creates conflicts. Like, I'm like, I'll be very transparent with you. Probably half of my business is brand deals with software companies and there's certain things they're not going to want me to say. And like, the reality is right now, I don't know, like what aspects of software go away in an era of age, AI agents being really good, I don't know what software that wipes out, if any, I don't know what software definitely remains, if any. I Mean, like, my North Star has always been. I work for accountants. Like, I work for accounting firms and that. And in any situation where that is in conflict with, you know, brand partners or something like that, like, we always have to take the account side. Like, that's how we got started. I started making content because I couldn't find anybody that was producing content for accounting firm owners that didn't have Intuit and Zero in their pocket, where it was like, the main way they made money was selling affiliate. And I was like, well, that kind of sucks. I don't actually know if I could trust anything they're saying. And, like, we have our own versions of this now where, like, people will find reasons not to trust what I say in our brand and all that. And a lot of it is like, folks that will just, like, pop in for a video and be like, wait, you have app recommendations, but you just took sponsorship money from that company? And I'm like, well, a little more to it than that. But they're out as soon as they left the comment, right? So I'm like, I've been thinking about this a lot for. For what we do because, boy, it feels like the ecosystem that just. The industry is set to change. But, you know, who's going to know about it first, honestly? And I've. I mean, I've been saying this from the very beginnings of AI stuff. People who are most plugged into this, I truly believe, are the folks that are listening to this podcast, sharing it with the folks they work with, and are just, like, being real about, like, okay, all this is happening. What do I do today? What do I do tomorrow? Because right now you got a lot of people that need your help. Sure doesn't feel like anything's drying up, right? So how are we investing in the future while simultaneously spinning the plates that we already have, spinning all the people that we have committed to help. But this whole other plate, which is a big moral responsibility, is all these people that I employ whose skills need to be, like, still relevant three years from now. How do I shepherd my business into this new era? I worry sometimes about talking too much AI stuff on this podcast because it's easy to get. Get burnt out on. I know I can get burned out on it. I just think next six months, next 12 months, there's going to be a lot happening. And for me, frankly, like, I gotta. I gotta keep seeing that as a privilege. The fact that, like, I get to sort of take people's hand and kind of show them, you know, all the noise, you Know what everybody's saying out here, like, look at this, look at this cool thing. It will do, but this is inappropriate and this is appropriate. Like, I think I feel very privileged that that is my job and gotta stay laser focused on the doing and the examples and not the conversation. So much of the, I mean, so much of what people talk about online is just like the only reason they even started this conversation was for engagement. They're not trying to move the conversation forward. And so it can be hard to get sucked down those rabbit holes. Social media has been really hard for me lately to get excited about, but going to be interesting six to 12 months. We're going to keep talking about it, but next. Boy, this feels I'm going to pull a hammy with this pivot. Next, let's get back to the stuff that like is sort of the cornerstone of an accounting firm is the clients that you serve. Heck of a pivot, right? Holy moly. That is what we talk about on this podcast is like strategy around running a better accounting firm. I don't normally subject you to 30 minutes of me yelling about AI. New segment just dropped, buddy. Niche of the week new, new graphics, new jingle to let her rip. Let's talk about the niche of the week. Not bad, right? Tell you what, we got a top notch, top notch production team around here. Niche of the week. This is where we look at a specific niche that a accounting firm could go into. And in most cases there already are accounting firms inside of that you may not have thought about, may not be aware of. You probably frankly have no interest in going into this niche. But by going through the exercise of digging into is this a good niche or not? What makes it a good niche? What are the steps that we would take to get into this niche by sort of you going through this exercise and maybe putting this coaching hat on to say, like, how would you coach someone to go into this niche? I think it makes us more aware of how we serve the clients we have today and how we have maybe more agency than we realize to go after a more specific type of client to create an easier and more profitable business for us to run. So we are looking at high end custom jewelry designers. Now of the 8 to 10,000 jewelry manufacturing companies in the US about 2 to 4,000 represent sort of the sort of custom bespoke designer subset of that. Now that is according to like industry surveys. I think in all likelihood this is understated because you have an entire. It's almost like accounting firms where you have a huge number of accounting firms that maybe don't even call themselves or accounting firms call themselves firms or jewelry designers that are like, I don't know if I can even call this jewelry design. They're like going out and making stuff and selling stuff. So there are thousands of these clients that you could potentially work with. Revenue range, there's sort of, there's a few bands here. You've got small sort of solo custom designers operating around 100 to 500 grand a year. Others with, with more of a direct consumer like their own website, usually their own social media and kind of their own fans doing a little bit more of a scale online selling thing that'll come in between 500 grand and 3 million a year. And then a step beyond that without getting into the humongous shops, you've got some bigger DTC like direct online jewelry folks doing 3 to 8 million a year. So one of those pockets is probably most interesting to you if you're like a long been like an entrepreneur, like solo hustler, like maybe that 100 to 500 grand, that sort of person that's getting into business for themselves for the first time, maybe that's attracted to you. One that jumps out to me is probably the 500k to 3 million group. They are outgrowing the solo thing, looking for a little more sophistication. They also have a little more money to spend on professional help. The 3 to 8 million group, that's interesting too. But if I'm looking to start somewhere, what I would probably get excited is that half million to three million band in terms of jewelry, like the market in general, it's projected to grow 8% a year through 2032. So there's nothing like super undermining it. The biggest shift within the industry is lab grown diamonds. Demand for that is up 55% in the last couple of years. But 60% of millennial and Gen Z jewelry buyers are now seeking customization options. So we've got some good signals here that like custom jewelry stuff is something that's probably here to stay and grow. I also think myself that there's like a. As mass production is kind of the norm, there's just a growing appetite for more craft and more taste being kind of intertwined into literally everything. And the age I think makes this an even bigger thing. And it probably just looks like talented people having their own little micro followings and they're making cool stuff that people are happy to buy, which is amazing. Most owners of these businesses are the designer owner, so they are, they are the pie maker, as it were, that goes into the business to sell a bunch of pies and they're like, what the heck is all this other stuff that I have to do now? How to find them in terms of associations. You got the Jewelers of America, founded in 1906, the largest US trade association for the jewelry industry. I don't know that, that one's interesting to me. That feels a little like old money. Other organizations, the mjsa, Manufacturing Jewelers and Suppliers of America, it's a trade alliance for jewelry making. You got the Society of North American Goldsmiths, the American Gem Society, Jewelers Vigilance Committee, which is more around legal compliance. United States Jewelry Council. I look through stuff like this and I'm like, is there a group that has the respect of like the demographic of the type of people that I want to serve this? You know, for a lot of folks this can feel like, you know, maybe the aicpa, like something that is a little bit more old money, that is not without value, but maybe isn't the thing that's getting the type of folks that I want to work with excited. Associations can be a great choke point to kind of find where a bunch of a certain type of person are collecting, but won't be, won't make sense in 100% of niches. It's that time of year, the time of year where your hosted tax software starts grinding to a just to a halt because everybody's trying to use their computers at the same time. This is because when you host your apps with most providers, like your tax software provider, the servers you're accessing, they're shared servers. Everybody and their mom are logging into those things, trying to get stuff done at the same time. The solution is today's sponsor, Verito Man. Verdo's gonna give you your own private server where all your tax stuff, even any hosting needs, QuickBooks, desktop, whatever you got, it's hosted on a server that only you're accessing. So it doesn't matter how many people are hammering that server, the only people hammering it are you. It's easier than it sounds to get started. They can actually transition you in just a weekend. And especially as we're heading into busy season gang, this is the time to do it. Bardo. Vastly cheaper than the host you get from your tax software provider and faster, like kind of a no brainer. Now Verito can also do your whisp, the little security thing you got to do. If you do tax work in the US they can also be your entire IT department. In fact, the cost of hosting your apps and them being your IT department is less than Thomson Reuters Virtual Office know a lot of people, they're very happy Verito users. In fact, the average response time from Barrow Support if you come to them with an issue is now less than than 60 seconds, 24 hours a day, seven days a week. Like I said, now's the time to do it before we get too far into busy season. To learn more about Verto, check out the link down in the show notes. Hey real quick. Some very cool new updates were just shipped by one of the leading practice management systems in my app recommendations this episode Sponsor Financial Sense. So if you do bookkeeping work or client accounting work, this is worth checking out. They're building some month end close automation features into the practice management system. So stuff like handling uncategorized transactions, flagging transactions without payees, without class, without location expenses and bills without attachments, inconsistent expense classification, large transactions. It will now flag all of these items inside of a month end close in Financial Sense because it's syncing the actual Ledger data from QuickBooks Online to Auto flag these items. Really cool since stuff that you've only seen to date really in like dedicated workflow tools. Financial Sense now, which is a practice management system is pulling some of the best of that workflow into the platform so you can do it all in one place. Pretty cool stuff. And I got a discount code for you. If you sign up for their team or scale annual plans with two or more users, you can use code Jason10 to get 10% off. That is financial sense doing some cool bookkeeping automation stuff. To learn more about them, check out the link down in the show notes. Examples of some upcoming events. You got JCK in Las Vegas. It's the largest jewelry trade event in North America. 30,000 plus attendees. That's happening at the end of May. We got Couture Jewelry Collection and conference also happening at the end of May. Oh it's right before that conference in Vegas. So you got those two back to back. Atlanta Jewelry show midway through May. Jewelry Jewelers International Showcase first couple weeks of March at the Miami Beach Convention Center. As I look at the details for this, this looks a little bit more like you're wanting to get discovered by brands or something like that. A little bit more of a sort of like the maker type person maybe showing up here rather than like the giant companies at the bigger more traditional events. A few others very similar to that. Always we're thinking like within a domain. Who are the People inside the domain that I'm most excited to hang out with, most excited to serve. Not unlike accounting events like there will be events, some events that those people love, other events where people are like, oh gosh, I would never touch that with a ten foot poll. Software, this is one of the easiest ways to get in front of a bunch of people within a space is to partner with the software companies who serve that space. If you look at say CrossFit gyms, you've got a pretty small number of platforms that gyms like run the entire business on. Imagine if you were like the person that was like partnering with them to create content and kind of the most visible accountant within that network. Crazy thing about software companies is venture capital will give them money that has to be spent. In fact, most estimate that 80% of like a series A round of financing goes into Google Ads, meta ads, all that stuff. That 80% of money's got to go somewhere and they are hungry for content. For people that will talk about what they do is a great chance for you and your accounting firm to be able to talk about how their software solves valuable problems for your clients. And just by going through and doing that with them, they're spending their marketing dollars to put your identity on blast. Now for jewelry. I don't know if this that actually maps very well to this industry. Few software companies here Matrix Gold, It's a professional 3D CAD software for jewelry design starts at 6,000 bucks for new users. I don't know if like the folks within the revenue bands that I'm looking at are using like CAD design for making jewelry. Maybe they are Countersketch, another Software that's similar. Rhino 3D is a similar one but a little lower cost. And then you've got a few different point of sale like retail jewelry specific options. You've got the edge. You've got jewel360 Ah, this one though carat IQ. This is more an accounting and inventory platform. Ooh. With a QuickBooks integration. This is a juicy one. If I'm going into jewelry, what I would be more excited about is serving the folks that are selling jewelry online rather than the in person retail stuff. Two very different businesses that probably just comes from my interests rather than one being better than the other. But Carrot IQ looks to be the system that connects to your like digital like online point of sale, not in person retail point of sale, but like your Shopify or something like that. And then the accounting and inventory management happens inside of it. Those especially when they hold themselves out as Being an accounting platform, yet they integrate with QuickBooks. Isn't that interesting? Those are the type of platforms that super want an accountant to come and make content with them. We've all seen bad versions of the partner referral directory from software companies. Like right, like the QuickBooks Pro Advisor is probably the best thing we ever had in the accounting industry and even that was limited. But so many other accounting software vendors are like, oh, we'll put you in our partner directory. And like you never get anything from this and rarely when you do, it's like some weirdo and you're like, I don't want to work with this person. Industry versions of that can be really good. Like I've like I know somebody that built an entire interior design practice literally just on the referrals they got from this one software company saying, here's our few preferred accounting providers for folks that use our platform. Like literally built their entire accounting firm on that. So getting in with a software company to be the preferred provider or on that shortlist for them, it's a great way to build a very niche client list. Consultants, always a good way to get into a space. And jewelry, you've got lady called Tracy Matthews who sounds like a sort of Jason Stats, but for jewelry type personality you've also got Larissa Worstwick, focuses on jewelry marketing consulting, newsletters and publications. You got a few like traditional looking magazines, also a couple like trade reports sort of things. Sometimes those are worthwhile, sometimes not. Influencers. We're back to Tracy Matthews. She's got a podcast, she's got an online community, major voice for jewelry business owners, buddy. You find somebody like that that's got an audience of people that are. This is the hard thing is a lot of times these podcasts are about like making the thing you want the podcast that's about the business of the thing. Like that's the best place for you to show up. The jewelry industry's got this couple other folks. Ben Baller, celebrity custom jeweler, is the host of a podcast. Another couple names. Not sure though if they're talking about like the doing of the jewelry versus the business of the doing of the jewelry. Online communities. Actually a surprising number of online communities for folks that are in the biz required training pipelines. This is a great place to get in front of folks. If everyone has to go through the same process to go into that type of business. And you can show up there and be the person for tax or the person for accounting for that type of person, you know more work than you can handle jewelry stuff. They have a few versions of this. There are like certifications that you can get, programs you can go through. Some states have or most states have certain permitting that's required for jewelry. So stuff around precious metals, jewels, pewter, interesting, specific specifically around the marketing claims that you can make how you communicate the metal content. This is actually an interesting angle here. So standards around how you have to report the metal content and the jewelry that you sell. There are some fincen anti money laundering rules. So dealers exceeding $50,000 a year in precious metals or stone purchases and 50,000 in gross sales must establish written anti money laundering programs and file currency transaction reports. What, what I'm seeing is niche specific compliance stuff which is gold is awesome. It is a great way for you to demonstrate. We always go back to like the accounting firm website where you could say we're an accounting firm for jewelers. Great. Better than nothing. But how about we can help you do X, Y, Z and like literally getting into a specific example with the vernacular where the person's like oh dang, like they're not saying it, they actually get it. Compliance that's specific to niches that you can help with. That's a great way in to finding those folks. Specialty insurance companies can sometimes also be an interesting co marketing angle. Not sure there's much of that in the insurance space. There's like one company that's humongous. Got a few dominant suppliers. Nothing in the way of group purchasing organizations. Pain like accounting pain points though that are specific to these types of folks. Similar to compliance pain points that are specific to a type of entry. These are a good opportunity revenue recognition for folks doing like custom commissions. So custom pieces usually usually require a 50% deposit up front. Proper revenue recognition requires tracking work in progress against deposits received. Maybe a threshold where like that's an expensive problem versus it isn't. Like if your average project is a hundred grand. Like yeah, that stuff probably matters if it's itty bitty maybe not consignment very common in this niche. So inventory management, selling in multiple locations like doing pop ups. Oh, precious metal price fluctuation. Yikes. Multiple valuation methods. Fifo, lifo weighted average multi category inventory. You've got raw metals, loose stones. That's all like very inside baseball hard stuff. Right? Like these are the things that you're looking for when you're trying to find a niche where you can solve expensive rare problems. Who are the people that have rare problems so that if they just walked into another accounting firm and they're like, yeah, no, we can probably figure it out. It's not realistic that they could. But also they have the money to pay well for help. That's the combination. You're looking for rare problems and they got money to pay. If I had to grade this niche, high end custom jewelry designers, maybe B, maybe B minus. One thing that's a little tricky is you can see a lot of old money in jewelry. And at least for me, that wouldn't be the type of people that I would be excited to serve. I'd be excited to like serve the more the independent folks that are like, man, I just started selling stuff on maybe on Etsy at one point. It was a fun side thing. Now we're doing a half million a year. I got to build a team and I'm in over my head. That's the type of person that at least for me, would be exciting to serve. I'm gonna, I'm giving it a B or B minus because I'm not seeing here an abundance of choke points where everybody has to go through a specific process. And I could grab all those people there like Holy grail. An example we've talked about in the past is someone we talked to who just works with grocery stores of a specific chain. And everybody that's going to be like a franchisor of that grocery store and it's technically, it's not franchise, but you know what that means they have to go through this sort of tax and accounting course. And so they fly this person out twice a year to teach all the new people going through the pipeline. And they get 80% of those people by default. That's an example of a niche that's like, oh my gosh, if I can get into that choke point where everybody's coming through you just by default, you're going to be the person for them. Right? I don't see a super, super strong example of that in this niche, which is the only reason why I'm not giving it a bigger grade. But the North Star here is like, is, is if this is something that you're passionate about, like if this is something that you love, that's when you're going to be at your best, is when that Venn diagram of here's the things that I'm into and here's what I can actually see as a good niche. If those two things overlap, oftentimes there will be something that we love that isn't a great niche. There's not a person in that niche that has good money to pay for you to solve those rare problems. I hate to say it, a lot of times restaurants are this passionate about restaurants, but the type of restaurant you're working with doesn't make very good money. There are restaurants that make very good money, but not many of them, not most of them. So in a perfect world, and the holy trinity is, do they have rare pains? I can solve? Because not all industries do. Like some just they have very simple, they can be done cash basis, not a big deal. They don't have rare pains. But I want somebody that has rare pains. They have the money to pay for good support. And it's something I can get into, something that I enjoy. Like, if you don't want to talk to dentists, all dentists, all day, then the fact that dentists have money to pay is not enough of a reason to go into like niching into dentists. Right? Well, one thing I forgot to talk about is there's actually there's a few firms in this space already, which is totally fine. Is a great signal. In fact, that can be a great place to start is talking with those firms and being like, what do you not want? And this might inform your positioning a little bit depending on how visible these firms are. If these firms are at every single conference and all that already, it might be hard for you to come in. Like, you'll just have to have a response to that. Either you're serving a slightly different profile than they are, which is great because then they can refer to you, all the folks that like are of your profile. Or maybe you're coming in and their hand me downs are totally fine for you. Like if that firm has a $10,000 a year minimum and you're just starting out, awesome. Go like impress that firm, show up at the same conference as they are and they will start sending you clients hand over fist that are now their hand me downs but great clients for you. If you see absolutely nobody serving a space, it might be an opportunity. It might also be a signal that it's not a great space to go into. But don't be spooked by other people being in that space. I would still say the biggest error I see folks making when it comes to like getting a little more specialized in the type of client you want to serve is just not getting specific enough saying, my niche is going to be real estate. I was like, okay, real estate. What? Like even here we talked about jewelry. You could explain it as jewelry, but it's like, no, it started as high end custom jewelry designers and ended at custom Jewelry designers selling online that are between a half million and 3 million a year. Right. Like that's what we got to. So that's kind of a specific subset within jewelry. Most of us, believe it or not, that specificity actually makes it a little easier to find the people that you want to work with as opposed to serving everyone. Most of us are generalist out of necessity. Like that's the business that came to us is we're serving anyone and everyone. And honestly, it's a skill set to learn how to specialize and until you have that skill as a tricky thing to do. But that's why we're here having this conversation. In practice, specializing is not this burn the ships YOLO moment. It is a progression. It is a decision that the next client that I want looks like this. And it only ever gets to the point where we are a quote unquote niche firm when there's an undeniable amount of evidence that that's a more fun, more profitable type of client to serve. And then you got a super cool business, a profitable business, one that's easier to systematize because it's more focused, it's easier to hire people into as well because somebody that's into the same thing who is an accountant will be like, dang, a firm that does this, that's super cool. All sorts of cool advantages. A lot of folks, of course, that are like, I want to be a jack of all trades. I enjoy doing one of everything. Niching down is not the right answer. I will say most people you talk to who say niching down is not the right answer have not niche down. The people who have niche down, it's very, very, very rare that someone who has niche down has been like, yeah, I did it, it wasn't for me. And so it can be hard to turn off the voices sometimes of the folks who haven't done it that say it's not the right thing to do. Right. Okay, enough diamond stuff. Not diamond stuff, jewelry stuff. Let's get into the mailbag. We got a couple more AI things to close the loop on and some advice for some early and mid career accountants that are like, what makes sense for me. Next. Next. Mailbag. Mailbag. This segment is called Mailbag. Got a few AI ones I want to knock out here. First, which agent do you recommend to handle either basic bookkeeping and or tax organization. Here's the thing. Right now, it's not the most powerful agent, but ChatGPT agent mode is the most accounting firm ready agent. And I hope that Changes soon because I want Claude Cowork to be in more people's hands. It's so, so good. But if I had to make a decision today, I'm still putting Chat GPT, specifically the business plan, in the hands of my team. Couple ChatGPT specific problems any way to use ChatGPT to solve the problem of bank feeds in my QBO? Constantly disconnecting across all my clients. Amazing. In 2026 this is still a problem, right? Just the inconsistency of bank feeds. To my knowledge, there is not an intelligent email notification for this. Part of the problem is QuickBooks Online's infrastructure is like these are all just like single company accounts. Like QuickBooks was built for the small business owner. Then they bolted on qboa, which is like the layer that we access our client files from. It's the layer where you see the list of all of your clients and you click into which one you want to navigate. That layer is not very intelligent. Like there's not a way to like aggregate information across all of your different company files. Maybe that gets better with the new Intuit accountant suite stuff. I hope it does get better. It almost certainly will. But can we use ChatGPT to solve this problem of bank feeds constantly disconnecting? This is actually something that I've tested before is using ChatGPT agent mode to log into every single accounting file that I have access to and check the banking status of each. It can do that. There will be a maximum limit to it because an agent's only allowed to work for so long. So I'm not sure how many files it can work up to. I think I tested it on a QuickBooks account that had access to like 15 files and it did it fine. The problem is I wouldn't because I still don't trust an agent to go into an online environment where it has read write access. I don't trust an agent to do that unmonitored. Yet I can maybe trust an agent to go into an environment where it has read only access, like a research library, something like that. Like maybe I can trust it to do that stuff on the web when it just has read access. But the fact remains, in a QuickBooks file, all it takes is that misclicking something or getting confused and it could potentially break something. And so that feels like too high stakes of an environment to let it run autonomously in. So to bring this back to the question, how do I use ChatGPT to solve the bank feeds constantly disconnecting from QBO? The answer is you don't. I don't see a way to do that yet. I don't know if the QuickBooks, if you get an email from the individual files anytime a bank feed breaks. If so, I would probably try to set up a rule to auto route those notifications somewhere and then kind of centralize the handling of it. Right. Like, I've shared before how I had a pre accounting team. So the pre accounting team was responsible for logging in and getting all the bank statements every month and asking the clients any sort of recurring questions. Every month they got to give us X, Y, Z. I had a pre accounting team of admins that would gather all this information so that when the accounting professional came to work on the on the file, everything was already there. This is an example of where I would. If there are notifications for this, I can't remember, I would route that stuff to the pre accounting team so that they can handle it anytime those connections break. And it happens right away. Because the last thing you want is for your senior accountant to log into the QuickBooks file on the 10th of the month when the financials are due on the 12th of the month. And they're like, oh crap, the connection's been broken for the last 40 days. And now I need X, Y, Z. And your client's like, why didn't you tell me this weeks ago? Right. We've all been there. One other ChatGPT question. Is there any way for ChatGPT to solve for automating more transaction coding? This feels hard because there isn't like a standardized chart of accounts across all my different accounting files. Yeah. So how do you automate a bunch of transaction classifications when everybody's business is different, when everybody has a different chart of accounts, when one purchase from Costco might be 12 items that get categorized one way and then the next purchase From Costco is 12 items that may just be a bunch of hot dogs. This is why bookkeeping hasn't been automated, frankly. Like, there's. Just to properly classify a transaction requires more information than just what came through the bank statement. But in terms of like, how do I then automate that across a bunch of client accounting files? It's hard to do. The frustrating thing about accounting platforms is they've never given us a way to centrally manage bank rules across a bunch of client files. If you think about it, bank rules, which are kind of bank rules, are like how you automate bookkeeping still these days have a really good set of bank rules so that 80% of the transactions going through like they just are handled by the rule. But those then are unique to every single client file. I could have a hundred construction clients and every single one has a unique set of bank rules. Why does it need to be that way? It shouldn't be that way. At least with QBO you can export your bank rules and then import bank rules into another file. To my knowledge, Xero doesn't offer that ability. But that doesn't keep them in sync. So how do I have some sort of like centralized model for how everything across my clients gets classified? Even if I go through the work of putting them all on the same chart of accounts, for example, It's a hard problem to solve. If you look at other companies like Bench who was kind of one of the biggest automated bookkeeping companies in the US that was their secret sauce. They built their own custom accounting ledger from scratch and behind the scenes they had this, they had kind of centralized rules so that when an account, when a transaction came through for one small business and another small business classification could be automated across a whole bunch of different company files. We accounting firms, we don't have tools like that unfortunately. They'll let me like standardize bank rules across, you know, 500 different QuickBooks accounts. The closest thing I've seen to this is now what digits is doing. And to be clear, digits right now is still only good enough for cash basis bookkeeping. I say good enough. That's not the right way to say it. It's only really designed to handle fairly simple stuff. But they have sort of this layer cake analogy they use to describe classifications. They will very first look and see have there been other transactions like this one in this file before? If so, classify it the way that it was classified in the past. Second, they will look and see at the accounting firm level does the accounting firm have other similar clients where they've had transactions like this in other companies that they coded a certain way? If so, I'm going to use that to try to auto code this transaction. If that isn't there, then it looks across all digits bookkeeping vials and it says what do other people do? And then to the best of its ability it auto classifies accordingly. And they, they are intelligent in how they talk about this. They're like one of the kind of part of our secret sauce is that we're able to determine the similarity of different charts of accounts. And so if one company has one set of chart of accounts and another company has a totally different one, then when company a code something some way the platform is able to Kind of infer from that. Even though it has a totally different set of chart of accounts what the most similar account to that is and start with that as a recommendation. It's pretty intelligent stuff. I think they are pushing it as far as it can be pushed right now. I still don't think it's a thing that it's possible to fully automate right now. But we want to get it as far as we can, right? Hey, real quick, I want to invite you to my super excellent weekly newsletter. My newsletter was like the start of my thought leadership journey. It's where I try to collect what I think are the most important things for you to see every single week sent to you every Wednesday morning. Make sure you don't miss any of the good stuff. Right. Increasingly, we're also building into that some like of the best evergreen thought leadership I've ever done. Kind of make that like the most action packed, highest quality email you get to your inbox each week. Because listen, we're all busy, all right? You're not gonna see every single little thing that I put out. Good sakes. There's just too much. If you're not on my newsletter yet, check that out jasononfirms.com or I'll put a link to it down in the show notes. Hey, quick word from one of today's sponsors. Me. Me. Yes. I run a community realize that is for accounting firm owners and realize is kind of the secret behind how I've done all of my content and been able to like financially turn thought leadership into a business. About 100 firm owners in there realize is a systematic way to build connections with other peer firm owners. It's not a framework for how to run a firm. It's a way to introduce you to other super cool firm owners. Right now we're putting together small groups so you can have a group of like 5, 6, 7 other firm owners just like you that you meet with every single month run very similar firms. Having this for me, totally transformed firm running my North Star. Like my goal for you coming into this group is to make a new best friend, somebody that runs a firm. Because that was what I needed in my journey. The whole community is built around that. It's called realize link down in the show notes if you want to learn more. Next question. I thought this was a fun one. Do you think it's possible to start an accounting business where all you do is cleanups? All you do is cleanups. You believe that there's a certain type of accountant out there and this Must be one of them where they just love a good cleanup job. Rolling up your sleeves and really getting in there. I don't know. The thing is, cleanup means something different for everybody. For me, cleanup was a lot of plugin Cleanup was clients not going to pay me to go in and fix every single little detail thing. For other people, cleanup is like super ocd. Like who I get to go in and change every single one of these transactions to what it should be. That's kind of what makes cleanup work. Hard to charge for though, right? So is it possible to start an accounting business where all you do is cleanups? For sure, for sure. The hardest thing about that business would be how do I keep getting work? Because it's all project work, it's not recurring work. That is a great thing about accounting firms is your tax work. At a minimum, it's annual. You know, in Australia you have required quarterly business tax filings. And that's why accountants over there generally have a much tighter relationship with a small business. They have to fix the books and be in contact every single quarter. In the US it's once a year. And so the default for most tax firms is like, well, I'll just see you to do your tax return once a year. And the conversation doesn't go much further than that. I could see you building a sales engine from other accounting firms where they're like, we don't want to do the cleanup. Like, you clean it up. We'll take on the monthly ongoing work. Like if there's a hard nut to crack with building a firm that's just cleanups, it is, how do you keep continue acquiring customers? Because everything is project work, but you can do it. Nothing's impossible if you're connected to enough other firms. That's probably where I would start is try to find a bunch of other firms who would happily send you that cleanup work. I want to start my solo firm, but my job in forensic accounting and private equity feels stupid to leave. I'm making 200 grand with 4 weeks PTO and fully paid premium health care. So my question is, what do you think is reasonable first year net income for a solo firm doing CFO and business tax prep? I only want to work 40 hours and I have no client base already. Should I go after it or should I stick with my 9 to 5? I get a lot of kind of golden handcuffs questions where people are like, I do something and I'm paid so well for it, I can never leave or I do something that's so specific, I'm never going to be able to find another job where I can deploy this specific expertise. And I think we just underestimate how big the world is and how much opportunity is out there. I mean, the right answer here is very much dependent upon life circumstances, right? Like, do you have kids? Is taking a step backwards in lifestyle, Is that not realistic? But that could just mean that the right answer for you looks different. Starting with no client base, that's very hard. But if you look at like no Logan Graff, he bought a small book of business to cash Flow the start of his journey. So like that could be the option for this person is maybe I'm in a life situation where the cash flow really can't slow down. Maybe what makes sense then is going out and buying, buying a book of business that you can cash flow. Maybe the solution is just saving for a couple years. Maybe it's building the side hustle until that is getting close. It's just a question of like in your life circumstances right now, which of those is going to make the most sense for you. The most unhappy people oftentimes are the people that are getting paid really, really well doing a stupid specific thing that they feel like they can never leave because they're trapped. And the one thing we have to remember is at a hard life change like that all that matters is what's the downside and how do we make a plan for it. So if it doesn't work out, could you go back to doing forensic accounting and private equity? Probably. I mean most accountants, like that's the ultimate accountant cheat code is like something goes sideways, what are you going to do? You're going to go back to being an accountant and like that's, that's at least historically has not been hard to find a role being an accountant even at the same job that you left, like they're so desperate for help they'll be like, yes, we'll happily let you come back. Let me actually let me tie this into another question that is in this same block here because it's very related to this one. I'm a senior in college and I'm wrapping up my master's in accounting. I have a year of tax tax experience with my current firm and want to start preparing on my own this next tax season. Nothing complicated, just simple to start a returns. I really want to go for it on the side to start, but it feels like there's some large barriers to entry. People around me seem to think this isn't realistic or that it'll be too Much. What are your thoughts? Here's the thing, brother. Tax work is very, very hard. There will never. There's no number of years you can be in tax and then go out on your own and be like, yep, I know it all. But it is. Of all the things you can do within running an accounting firm, it's the hardest thing to go and bounce and sort of do on your own own. Now there are now a growing number of communities of, of folks that are solo tax folks that go out and do their own thing and they can kind of lean on each other's expertise and get more support. That definitely helps. But what you're describing here when you say people around me seem to think it isn't realistic or that it'll be too much. What you got to realize, and this applies to both this person and the person in forensic accounting, is when you are in that other world working for somebody else, nobody around you has done what you're talking about doing. So of course they'll be like, no, don't do that. Recognize that everybody around you will be in defense of their own life decisions. And so nobody's going to be like, yeah, no, that's definitely the thing for you to do. As opposed to if you went and sat in a room, people full of people who had just done that thing, they're all going to be like, oh yeah, man, this is the only path. So, like, there will never be a day when the people around you are like, no, no, you should definitely go do that thing that none of us are doing because they don't know anything about it. Like, they haven't done it. Like, it may not be interesting to them. Like, for sure, it's not for everybody. But the fact that people around you can't understand it is not a reason to not go and do it. Everybody in life is playing a game. And when you work inside of a company, everyone's trying to win the game that they see as being like the game to play inside of that little micro universe. And not enough people stop to consider, is this even a game worth playing? Like, like, what am I doing here? Is this the best game that I should be playing? Most of us will spend like decades of our life just heads down trying to win the game that we are in. And so at no point will any of the people around you who are heads down trying to beat you at that game say, you know this, yeah, you should definitely go and do that other thing that I don't know anything about and that you've thought about. Ten times more than I have. Like, no one around you will think that that is part of why. And this is the same advice I give to people who have already done this. This is part of why it's such a great life hack to surround yourself with people for whom their normal is what you want your normal to be. Hijack their normal, as it were. If you got people that already run accounting firms, you want to do that someday. Go hang out with people that run accounting firms and you'll just hear how they talk and before long, it won't be scary to you because it's just their normal. If you want to charge $10,000 for tax planning, go spend a bunch of time hanging out with people who charge $10,000 for tax planning and it becomes your normal. So if you are at work every day surrounded by people who don't do that, those are never going to be the people that, like, give you the confidence to go out and do it. Now, if it's something you really, really want to do. There's tons of people who listen to this podcast and hang out on Tax Pros, Discord and like, all sorts of stuff that are in the process of starting their own firm. Go and hang out with those folks, hear their stories. That will either give you confidence or you'll see more kind of behind the curtains and be like, yeah, maybe I don't want that actually. But that at least is educating you and moving you forward. But it will never be the people around you in your work environment that will give you the confidence to go out and do it. The biggest blocker, like the practical blocker here is one year's experience going out and launching your tax firm really, really hard. Frankly, at one year, like, you're hardly doing anything more than people ought to be doing in TurboTax. Probably took me seven years to feel confident being able to talk through 90% of small business and personal tax stuff. But even at seven years, I knew nothing. I still felt like a total idiot. So tax is hard. Tax is hard in that way. But I would say there's also probably a little bit too much of people like me, people with loads of tax expertise that love to pull the ladder up behind them, right? And be like, well, anybody that doesn't know as much as I know has no business of going and doing this stuff. It just takes an awareness of, like, where are my boundaries? Which hard. Sometimes you don't. You don't know which don't know. And then you will have people that will do that recklessly and they can kind of become a target and people will be like, see, like that. That's why you shouldn't do it as folks like that there is an intelligent way to do it. Another factor here is like, think about it in terms of what's the fastest way to grow me? Like, what's the fastest way to grow my capabilities? At different stages in your career, that answer is different. Being on someone else's payroll for a decade and them paying you to learn, oftentimes that's the fastest way to learn. But you're not learning everything, right? If ultimately your goal is to run your own business, you are learning one set of technical things, working for somebody else. When you go start your own business, you're kind of learning a different set of skills, right? I think, honestly, I think the North Star for most people is, what am I going to be really excited to get out of bed doing in the morning? Like, I feel incredibly privileged that right now I shoot out of bed and see what did my cloud bot send me overnight and think, how do I build this into a video that people will actually watch? And if I go back to my decision to sell my firm, it was I could either go out and make this content, which I kind of enjoy and I feel like I'm actually helping a lot of people with it, or I could go back to the firm, which I also enjoy. But it's taking a huge amount of my time and it's taking it away from my time with my three very young kids. And so at that stage of my life, there was one thing that I was shooting out of bed to do. And oftentimes I think, like, you have to be careful because sometimes that can be fleeting. But if you know you're going to do something, I think the answer is usually do it sooner than later because oftentimes people do the opposite and they wish they had done it earlier. Two more to go here. One on three tier proposal optimization. They've been doing three tier proposals where each service, tax bookkeeping, tax planning, etc. Has its own dollar amount listed as a line item. Do I even need to read the rest of this question? So all of the items on the proposals have their own prices. Each it's itemized. For each item, you probably already know where this is going to go. Back to the question. So when clients look at silver or gold, they can see exactly what each individual service costs and how the tiers build on each other. The problem is this gives clients a menu to shop from. They'll say, I would like A bronze. But can you add tax planning from silver? I don't need the other stuff, just charge me bronze plus that one line item. They're basically building their own custom package using my prices against me. And I feel stuck because the numbers are right there. I'm realizing I should just price each tier as a single flat amount. Gold is X, silver is Y, browns is Z and list what's included without dollar amounts next to each service. Yep. What's the question? You answered your question. Yes, this is the problem. Menu based pricing. I will take this one step further to say yes, stop putting prices on every single individual item. Otherwise this is what clients will inevitably do. But also take it a step further. I believe that it is not the client's place. This is a hot take to decide the scope, that is how much work we're going to do for them. It is too easy to do a really lazy client discovery call where we're like, yeah, we can do this, we can do that, we can do this. And I'll send you a proposal, all those different options and you can decide those. They don't know anything. They don't know what is worth having you do or not do. Like you know that better than they do. And the easy example of this is tax planning. Like most people have never had tax planning before. And you will put a proposal in front of them with no tax planning, some tax planning and all the tax planning. Those are the three options. And they're like, I've never had tax planning in my life. I honestly don't even know what that word means. So I don't know if that money is well spent for me or not. Do you know? And in all likelihood you can be like, yeah, actually I do know. With what you've got going on, this level would probably be appropriate. Like at this level, I'm pretty sure even though you're paying me more, we can make it still roi, basically. And so the way I build three tier proposals is the scope is the same on all three tiers because who knows better how the scope should be set than you. Now if we do a really lazy version of discovery, we may not know. And so that's why we got to have a deeper conversation around like, who does your payroll right now? Who does your bill pay? Oh, you do your bill pay, you're signing checks and you're the business owner. Tell me what other valuable ways could your time be used? Okay, that's probably a great trade off for you to then pay me to do to help manage your Bill pay, because you probably don't want to hire another person just to do that. And we can do it in a more trusted way and you can still approve it digitally. You have that deeper conversation and you can see the opportunities in their business that better frame your value and then you pull that into the proposal. So the scope is the same on bronze, silver and gold. What changes from bronze to silver and gold is the service delivery. Do they have access to the big boss? Do they only have access to the staff? Will we get back to them in a week? Well, we only get back to them. And when we get back to them 24 hours a day, do they get your cell phone number? Do they have to call the front office or schedule a call? Something like that. So the scope is the same on all three tiers, but those tiers represent coach, business class and first class. And the great news about this change is when your three tiers are increasing amounts of scope, they're just paying you more to do more work. And I would argue that's not better. But when your three tiers are a more premium service delivery, then you're getting, I mean, at gold, you're getting paid over twice as much to do the same work with a greater level of access. That is a more fun business to build. Right now. If you had clients, if you had, if you had half as many clients and they were paying you twice as much and you got to work deeper with them, most people would say that actually sounds fun. That's actually something that I want. So I wouldn't let the scope be any different on those tiers. You decide what the scope should be. And yes, definitely don't put the prices on every single item within that proposal. But you're learning, bro. But you wouldn't, you wouldn't be here had you not tried this and sort of learned that lesson. Right? Last question. I love to see a POD or even just a Q and A session with you. Just therapy the accounting profession through the bipolar nature of the AI space. I get literal hot flashes with the ups and downs. One minute I can't count to 20 and they're going bankrupt, the next minute they're replacing 90% of white collar jobs in five years. Check swatch. Six months, check watch. Oh, nope, it's back to 50% of jobs by 2030 again. Yeah, I mean, that's, that's the nature of the news cycle, right? And people publications needing to have a headline that's worth clicking on every two hours. This is probably an info diet problem. There's so much value in curating our info diet and what we consume. Getting away from algorithmic feeds. I mean, that's how this all got started. That's. That's the nature of Clickbait is the reason I have to title a YouTube video a stupid way is that most people consume media from a feed of options, and more people will click on the hot girl than the picture of the spreadsheet. Right? So Clickbait started because people would go to a feed and be like, feed, give me a thing to consume. And I hope society starts moving back away from that. Podcasts are an example of something that's not that it's an RSS feed. Every episode I publish, it comes through, and you can choose whether to watch it or not. Email another example of this. Send you a newsletter, you can just decide, do I want to be on that newsletter or not? But every email I get, I see, and hopefully that doesn't go away. Like, they're trying to do the same thing with email by overlaying AI and email and just, like, serving up a summary of your emails. It's like, no. Email is like one of the last platforms where I can actually control what comes through the feed. I want to spend time. I don't want to kill time. I think every moment killed when I look back at the end of my life will be a moment that I regret. But I'm willing to spend time learning, I'm willing to spend time consuming from my favorite people that I look up to and respect. And so how do you curate your info diet around accounting AI? There's not a ton of great sources these days, honestly. Chad Davis on LinkedIn does a great job. We talk a lot about it here. Otherwise, the news stories, they're always going to be just that. And then you have the added problem of, like, trying to figure out what window of bias that communication's coming through. And, you know, it's coming from a software company. We went through Carbon's AI report that they do every single year, and there's a lot of really good stuff in there. But then there were a few things where I'm like, this feels like something that's like, this is kind of the Carbon angle version of that story. And maybe they're entitled to that because they. It's their. Their survey, right? They did that whole report, and I. I'm the same way. Like, Right. Like, I've got my own sort of perspective and views of the world and how it all works. And when I talk about this stuff, it comes through that lens as well. But I think the solution to what you're describing is better curating, where you're learning about that stuff and not maybe plugging into feeds quite so much, like not letting an algorithm decide what you should read. I'm increasingly moving away from that myself. And it's hard. Like there's a very real addiction element of addiction to that. But it feels really nice to take back control of what you're consuming. And you just, you look back and you're like, I don't regret actually not having spent any of that time on that platform. Platform. I for one, I'm very proud. I've never gotten into vertical video, like tick tock, Instagram reels, stuff like that. Like the nature of that monster. I can just see almost every one of those minutes I would come to regret spending on those platforms. And so I'm very proud. I've never gotten sucked down that don't kill time, spend time. And if you got folks that are really good on the AI education side, put that stuff in the comments. That would probably help a lot of us out. Just like cutting through the hyperbole. Tell me what I need to know. We try to do this on the podcast. We try to do it on the YouTube channel as well. If there's a tool that you need to be using, there will be a video about it. It's not a sponsored thing. It's not anything like that. It is like we have the world's most watched chat GPT training for accounting firms because I think it's essential. So we'll keep putting out content around this to like, make sure that you are aware. But from what I'm saying, like, it sounds like maybe just some of those channels, some of those feeds need to be turned off. Gonna be an interesting next couple year. What next couple of year? Cool it, Dayquill boy gang, appreciate you. Thank you for being here. You got any questions? Put that stuff down in the podcast. We're going to be back. We'll do this again soon and I'll see you there.