602 How Accounting Firms Pick Software Just Changed

Hey, Jason here. Before we get into the episode, we are just 2 weeks from the beginning of my 20-city live tour in-person events. These events we designed from scratch to be just for accounting firms. They are super fun. You'll meet a whole bunch of local accountants who also listen to the podcast. You get some continuing education. It's a bunch of snacks, an optional dinner. It is the only event I've seen where you can get continuing education credits with a drink in your hand. I want to see you there. It's happening in May and June. Did I mention this? Over 750 accountants have already signed up. I want to see you out there. To see the tour dates and get tickets for you and your team, head over to JasonOnFirms.com/events. Enough about your workshops. Let's get on to the actual episode. It is an exciting time to be an accountant as more and more accounting firms are finding ways to use AI agents in like actual client-sensitive workflows. It's kicking off the conversation around just how much AI agents are going to change our tech stacks. And this is all happening at like peak tech selection season. In most areas of the world, we're over our like first part of the year hump where you've got this workflow crunch and we're now like in, in what is sort of the Goldilocks time of the year of choosing a tech stack. So everything is settled before you get to the end of the year and you head into battle again. Like now is the time to do that. But how do I choose what tech to buy when like AI agents can seemingly do a lot. Like we're still figuring out what can't they do. They're literally doing like real, like production work for us now. So how do we make tech decisions today in that light? We're going to get into it. Plus, I'll walk you through building your first client-facing AI assistant. We'll talk about whether you should be sending out 12-month contracts to your clients now, not at the end of the year, but now. And hot on the heels of launching an AI to fix your QuickBooks, Perplexity just launched a desktop AI assistant that will control your entire computer. Should we be paying attention to that since they also launched an AI that would presumably prepare your tax return? And we're going to look at a firm that somebody just purchased and help them navigate how to best market an accounting firm in a 2026 that I think is going to go down as a big one for accounting firms. We got a lot of ground to cover. Let's do it, friends. If you're new here, welcome along to the Jason Aum Firms Podcast. It is a glorious week to work in an accounting firm, and those are the people that come and listen to this podcast. Accounting firm folk, if you're a leader, if you just work at one, if you aspire to have your own one day, what a dream. Daddy, I think I have decided what I want to be when I grow up. Yes, child. I would like to be an accountant. Okay. Why? Because I aspire to one day own an accounting firm. Honey, I think we need to talk about Lil Gibson. This is the only podcast you can listen to where that's cool. All right, you're in a safe space. Now, 2026, it is the year of the AI agent. It really is. We've had agents before. They're actually useful now. And inside of accounting firms, they are doing some wild stuff. Like, this isn't hyperbole. If you've seen my walkthrough of today-ready, like, use cases for Cowork on YouTube, Cloud Cowork can help you do everything from, like, actually get bookkeeping out the door to even, like, do real tax prep. The capabilities right now are outrunning our ability to meaningfully and responsibly implement them. But at the same time, you've got software salespeople kicking down your door, sending you muffins, calling you, "Hey, just following up because it sounded like you were interested in what we were selling. This is the time of the year when we're usually making those decisions." But how do you do that when you don't know if Cloud Cowork is about to ship an update that replaces that thing that you just like onboarded your whole team to. And so that's what I want to cruise through is actually 8 different changes to how we buy software in the light of what AI agents are doing this year. And the first thing that might have already given you like a tinge of fear and, and how I said that is, ah, we probably shouldn't be locking ourselves into long-term contracts. And there's a lot of software vendors that still do this where you have to come in with a set number of, I'll at least buy this much and I will buy X amount over 3 years. What in the world is 3 years from now gonna look like? This might be a knitting podcast by then. Probably not. But like, what will accounting tech stacks even look like 3 years from now? I have absolutely no idea, but I'm gonna feel like a real numpty if in 36 months I'm cutting another check for something that 3 years ago was relevant, but 6 months later was like, oh no, you didn't know Jason Staats? He actually just made a video on YouTube about how you can just do that with Cowork. You can swipe the prompt. It's not even hard. So right now, if somebody asks you to commit to a contract to buy software, ah, you don't want to think real long and hard about that because a lot of stuff is changing and it is changing fast. Second net new consideration is the interoperability— what a word— of our software. Our software's ability to work with other stuff has never been more important than it is now, specifically in its ability to work with your AI agent. If you are like me, you are spending a lot of time with Claude right now. And eventually it becomes an annoyance when you have to leave one app to jump over to another app to do something in it. We accountants, we know this all too well with the amount of app hopping that we have to do. And so the better side app for a job is the one that connects with your AI agent. So your AI agent can actually communicate with it and you can handle all that stuff kind of an all-in-one place. And so software companies are having to struggle with like, well, if I open up access to AI agents, like if we launch, you know, an MCP server is, is what this looks like in most cases. If we launch this and people stop logging into our platform because the AI agent does all that work, is that risky for us? Or even if it removes 50% of logins, is that risky for us? What does that look like when we actually do like user-based pricing, but a lot of users are now just using our platform from our AI agents? Like software companies are having to think long and hard about this. There's a lot of people that think 80% of software is about to go away and that Unfortunately, that's a possibility because if the AI agent can just do it, you don't need the other thing, right? And we're already starting to see that in some specific examples. I mean, look at document extraction. What was the company that Sage bought? AutoEntry. AutoEntry, you pay like per, you like buy credits and it'll process stuff out of bank statements and receipts and all that. Nobody's ever gonna pay for that again. Like that's such a trivial thing for AI to do now. And so like the software companies that you are investing time and money into, they need to be pretty forward on this stuff. We talked about an example of, of Double, a month-end close automation platform that now has MCP support. So there's a cloud connector that you can, you can add to connect into your Double account, but also on the Double side, and this is really important, the well-implemented version of this needs to give you like service accounts that are properly scoped. So I need to have control over how much the agent can do inside of that platform. Double does some of this where On a more granular level, you can choose if it has read or write access. But like a, a half-baked version of this is you just authenticating into, say, your practice management system, and the agent is able to do anything that you can do. Like that, that is not at all what you want. You wanna be able to lock down exactly what the AI can do, and its usage of that platform needs to be different than your usage of that platform. Like if the platform says Jason did XYZ, I need to know, did I do that, or actually did my agent to do that. And this is where quote unquote service accounts come into play. The way that my agent interfaces with that platform should not just be the same as any human user interfaces with it. And, and importantly, they're like, we want to be able to lock it down. Like, to, to give it access in the same way that I have access is like taking a desktop AI agent that can do anything on a computer and plopping it on my own computer, which is like, has admin rights to this stuff, and my password manager and Everything. I would never do that. A desktop AI agent, I'm gonna put it in like a sanitized environment, like on its own desktop where it just has like children's toys, like very, very, very basic things so that it can't get itself in trouble. And so the implementation of how my platform connects with AI agents, the, the details matter there so that when agents can actually do meaningful work in there, not just read access, but write access too. I can still dial it into only the things that it ought to have access to. A good example of that now is like in my media business, around the podcast and all that, we use Notion really heavily and Claude Cowork has a great Notion connector, but there's a bunch of nuance in the things you can give it access to, whether it has read access, write access. In some cases I wanted to have write access. In a lot of cases I just wanted to have read access. That connectivity with your AI agent and the other tools that we use, it is not a nice to have, it is a must have going forward. It really is. Consideration number 3 here and how buying tech is changing. Switching costs are going down real quick. If you have done a practice management system migration in the last few years, you know just what a joy that is, especially when the sales reps are like, oh yeah, what platform you using? Oh yeah, we can migrate you. We do this with people all the time. And then what that process actually looks like is somebody sends you like a spreadsheet template and you're like, okay, are you going to do this or am I going to do this? And it turns out you're actually going to do a lot of it. And then they import it and you're like, okay, this is like 50% of the story. Where's all the other stuff that I needed in here? It is a mess. It is such a big project. I did this several times in my 1800-client firm and it was a hoot. But imagine, like if you've had a big hairy project you've had to do lately and you try chucking that into Cowork. So often, like it feels like magic how good and how fast the result comes back. Like something big where you're like, how in the world did you just do— like I think I shared an example, like there's a software founder dinner that I'm hosting this week, uh, and I put my entire app recommendation database into Cowork and I said, it's like 160 apps in there or something. And I'm like, find all the founders that are in the New York area. And it came back in like 90 seconds and gave me a list of all these founders for 40+ companies, but actually looked into every single one of them. It came back so fast that I was literally like, I don't believe you. Did you look into everybody or did you just look into some and then give up? And it's like, no, like here's the full list. And where they all live. And it was so bloody fast. You can imagine having a humongous migration, like changing practice management systems where the system you're switching to has an MCP. You can imagine how like that migration that maybe at one point took 3 weeks could take an afternoon with like good connectivity between the tools and an AI agent. Think about this. I like, we are getting into a world where there's this AI agent layer of software on top of all the other tools that we've historically used. Like beneath that agent layer is all the tools that we've normally used every day from your tax software to QuickBooks, all these different things, way too many apps, right? But now we're almost abstracting that up to, I'm now talking with an AI agent that can talk to as many of those tools as possible. And as I say this, like for some folks this will feel pretty sci-fi. Other people are like, I'm already doing this and it's a real inconvenience anytime I have to actually sidestep Claude and go to another system that doesn't talk to it. But when we're doing most of our work with the agent, the tools that sit behind the agent, they don't matter as much as they once did. And so the notion of switching from A to B, especially with the agent's help, that's getting easier by the moment. And so this platform lock-in, like the really hard aspects of that, that's going away, which is, which is a good thing for buyers. It's a good thing for us. It's a bad thing for software companies, but it means we have more optionality. Consideration number 4. We have to care about audit logs more than we used to. So in big firms and other industries, audit logs for software are a really big thing, especially software platforms that will make automated decisions. Like, think about like automated trading platforms. You want to have a very detailed understanding of how did it get to that decision and what is every single thing that the software did on its own. Like, you need very detailed logging there so that a human can go back and review it if they need to, right? But audit logging inside of small accounting firms has really never been a thing. Like, we think about it through the context of the accounting ledger, and that's a very important place to have audit logs because you have any number of fingers getting in there and paying bills and making decisions and all that. And so you need to see who is doing what. But when it comes to the tools that we use to run our accounting firms, almost none of them support audit logging still. Like, some do on certain parts of the products. But like not at all in a comprehensive way like you would normally see from an ERP or a more sophisticated accounting ledger. And on the one hand, you can make the argument that like it's actually not a huge deal because at the end of the day, when work goes out the door, it's the responsibilities of the humans in the accounting firm to certify it, right? To say, yes, we reviewed this. It's correct. It's only going out the door because we signed off on it. So then why do I need an audit log from the software around what happened? But here's the counterargument to that is even if the humans are signing off on the work that goes out the door, what about when the software has a whoopsie and you find out about, about it? But even worse, what if it has a whoopsie that you don't even know about? I can tell you right now in using Claude, like the example I just gave you of it, it grabs this humongous Excel file, does a whole ton of research, and then just comes back. Ta-da! Here's the answer. I have got no freaking clue what it did on the backend to get all of that stuff done, which means if it has access to sensitive information and it's banging around on the internet and doing all these things, then just comes back with a response. How do you know that something problematic didn't just happen for it to get to that conclusion? Like, what if, what if BlueJay Tax research platform, what if one day they just stand up a landing page where anybody can go and ask a tax question? In fact, Marble, another research platform, they have this. Where you can just go ask a tax question. What if I give my agent some sort of tax task and it's unsure of itself and it goes out and sees, oh, like, here's a tax database. I can actually go ask this research question here. And it takes the client's name and like very client-specific things and it goes and it plugs it into this public landing page and it gets a response back. But then all you see is inside the AI assistant, It filled out a work paper or something like that, right? Audit logging, like it, it matters more than it used to. And so in the software that we use, if we go back to the example of Double, who recently launched an MCP so that an AI agent can talk to Double. If I'm gonna trust my agent with doing things, things inside of the software, what if it does something that maybe we don't intend, good or bad? But then we come to Double later and we're like, what is like, what exactly is happening here? We kind of need to see, right? And that's not work going out the door, but that's understanding like, I see this in our systems, where did this come from? Was it Steve again? Or was it actually my AI agent that did something that I didn't expect? I would say in the past, audit trails in software that we used to run accounting firms, general ledger aside, bill pay software aside, payroll aside, all that, like our internal tools, audit logs haven't been that big of a deal. But now because we're gonna rely on more software, more agentic stuff to just get things done, Those audit trails, they do become a bigger deal. And if we look at this through the lens of Claude Cowork, Anthropic, the company behind Claude, doesn't offer audit logs for the things you do with Cowork because that work is happening on your local machine, which is different than what they offer for chat, like the cloud-based part of their product. But to do an audit log for what's happening on everybody's machine requires an entirely different level of surveillance from them to see what you're actually doing on the machine and all the files that it has access to. And so you just, you fundamentally don't have that audit logging inside of Cloud Cowork, which is working on your local computer, which makes it all the more important that I need it then on the tools that it is interacting with. Okay. Consideration number 5, I would say the, the strength in the all-in-one versus best of breed debate, I would say there's now more nuance in those considerations. Do I try to go with one platform that'll do as much as possible for me, recognizing there's some trade-offs there? Or do I use 7 different apps that are all like the very best at the specific jobs I need done? In some ways, the ceiling is still higher if you're picking best of breed. And now your like horizontal AI agent, your agent that works across all those apps might make your life easier. Like it might ultimately be able to become the integrator that we've always wanted with the headache of these separate systems being like, oh, I gotta get this data from A to B. Like maybe AI can help us with that. But as we've been discussing, we know just how hard that is to do responsibly and securely when you've got an AI agent that is like working on the web and just doing a whole bunch of stuff and moving sensitive client information from A to B. Quick word from one of this episode's sponsors. If you are getting, uh, warmer to the idea of offshore staffing, as many accounting firms are, something that I did in my own firm and had tremendous luck with it. Wouldn't it be nice if you, you had someone you could trust to go out and find you the right person? You just pay him a flat fee to find you a great hire, like no middleman, no ongoing charges. Wouldn't that be nice? If you're looking for that person, look no further, look no further than TeamUp. TeamUp, run by actually a former bookkeeping firm owner who had tremendous success hiring folks in the Philippines, hired some local recruiters there. Now all they do is help place awesome accountants, be it bookkeepers, tax pros, virtual assistants. They source awesome local talent in the Philippines and they connect them with you. If you are into the direct hiring route when it comes to building an offshore team, that's TeamUp's sweet spot. They find you really good candidates. You hop on the interview and you don't move forward until you've got somebody that you're absolutely stoked to have them come and work with your accounting firm. If that's you, if you need some, some humans to augment your AIs, Get on the horn with TeamUp. More information down in the show notes. If tax season just punched you right in the grief bone, I have good news, my friend. There's a better way. We talk a lot about AI on this podcast and the group that is using AI to automate a whole bunch of the 1040 process. It is, it is one of this episode's sponsors, Soroban. With every like new headline you see around new AI models and new AI this, new AI that, the companies that are moving fastest using AI, they get better with every new AI model. And especially when it comes to the 1040 prep process, like if you're a US tax firm, oh my gosh, is there so much tedium still associated with collecting information from clients, getting it into the tax software, delivering it. It is a nightmare. If you are a firm doing like volume 1040 work, say 300, 500 1040s a year plus, this is a no-brainer. Take a look at Sorban. We've actually done some demos of them on my YouTube channel. You can even see how it works. Definitely worth checking out. More information down in the show notes. On the flip side, this is maybe the strongest argument for all-in-ones now, is the idea that if you have as much as your firm— of your firm as possible inside of one platform, then you can have an AI agent inside of that platform that's very easy to roll out because it only exists inside of that environment and has access to all of that information. And so that is, you know, a big all-in-one practice management system shipping a built-in AI agent where you can now use that agent inside of the platform to do all that stuff. Now, is it going to be as good as Cloud Cowork? I mean, it just won't because the companies inside of our space in the accounting vertical, like, they don't have the expertise, the same resources as an Anthropic to build this incredibly useful generalist AI agent. But it is infinitely easier to roll out because all it has access to is stuff inside of that bubble, inside of that system that hopefully houses as much information about your accounting firm as possible. And so if the notion of like ever rolling out, say, a Claude co-work, uh, across your organization to do like real client work, if that just feels impossible to you, and I mean, there's so much nuance and what are the right and wrong ways to do that. We got deeper into that on the last podcast if you're interested in that discussion. But if that just feels impossible to you, that might bring more people back to all-in-one because you can have that AI agent inside of a closed system. Consideration number 6. I think labor arbitrage changes. The notion of hiring a person in a lower-cost country to do a job that a person in a high-cost country could otherwise do. The question of, can I get enough cost savings hiring offshore to justify the trade-offs that come with that person not being onshore? Honestly, when it comes to production and getting work done, I think it's kind of just becoming more a question of who can put AI to work and who can't. Like, AI to me is almost feeling like a leveler. And so we still have all the client management aspects, the client relationship stuff, which historically have been managed by onshore people and probably will continue to be. But when you look at production, how the work gets done, how it gets to a point where it can go to delivery and be handed off to the relationship manager, how that stuff gets done, I mean, that, that's almost all being changed by AI, right? And because I think we all have so much leverage at our fingertips, like you can, very easily see a world where a person with AI can be twice as productive as a person who's not using AI at all. Like if you've seen much AI agent stuff, Claude Cowork stuff, that frankly feels very conservative. Like someone that is a wizard with that stuff today, it feels like can be many times as capable as the person who is not using AI. But that lever being so massive now, to me, like sort of levels the global playing field to like who has that lever and who doesn't. Because if you've got somebody that's an AI expert, a real guru, and they're offshore, aren't you gonna want that person whether they're onshore or not? Like if I am putting together, say, um, candidates to go and work for accounting firms or anything else right now of like offshore folks, I am pouring AI education into those folks like crazy. Because look at all the struggles we're having with upskilling our onshore teams. Like, there's a world where there's a country— you look at the Philippines, you look at a bunch of markets right now where a huge amount of that country's GDP is now supporting larger Western countries. Regardless of how you feel about that, there's a world where one of those countries goes all in on educating the workforce on AI, and you have these incredibly capable people that have all these skills that you can't seem to, uh, organizationally adopt. Like, get the folks on your team to really lean into, either because they don't have the time to do it, or we were just raised in a different age when doing this work looked so different that it's really hard to get our heads around how to use AI for this stuff. You can see a world where you end up with this labor force that is very, very AI capable and can just get superhuman amounts of things done. And because AI is such a big lever, I think the conversation shifts more to what can this person get done with AI than it does where are they. Where do they live? Consideration number 7. I could also see— so that's offshoring, like having folks that work for you offshore. I could also see outsourcing making a comeback. I've never been a big outsourcing guy where you hire a company to do 100 month-end closes for you and they're responsible for hiring and training the people and all of that. And at the end of the day, they're just sending you the output, a tax return, a month-end close. You could see a world where those companies, if they can nail the AI processes in a way that you can't, could become more compelling. Honestly, I mean, this is like what Botkeeper used to be. The whole notion that like we're gonna use a whole bunch of automation and RPA, robotic process automation, to automate, you know, QuickBooks month-end closes. Basis, and in some ways has evolved into this where they're building AI agents for more large firms, more ERP stuff. You can see a world where outsourcing to like very AI-capable groups, even, I mean, this could even be you. I mean, it could be accounting firm people right now who are like, We think we've nailed this. Now we're gonna go into the business of doing this for other firms. By the way, there's a bunch of firms that do this, that do outsource bookkeeping for other firms. There's an AI angle of this probably where this AI is a very hard thing for us to all collectively figure out, and it might be an argument for outsourcing to a more centralized way of getting this work done. Number 8, last one to be thinking about here, and this one's a little bit further out, which in this day and age means 6 months instead of 2 months. Let's not forget about our clients' agents, your clients who are using CloudCowork. Who's going to build the first cloud connector at the client portal level? The interface that your client uses to work with your firm, the client portal. Imagine if the client has a connector and can interact and push files and pull invoices and all that stuff from the portal that they use to interact with your accounting firm. Surely somebody is working on this at this point because it's just a matter of time before our agents, have to be able to talk to their agents. I used to give this anecdote when I was giving AI talks and more the early days of AI, where it's like, we have these tools now that will write our emails with AI. What happens the day that they get good enough to just send those emails out without even being reviewed? And they're like, you're like, well, the client's going to have that on their side too. So at a certain point, it's just going to be their AI agent. Sending emails back to me. And then our AIs will just be talking back and forth with each other and I'm not even doing anything. And like, what are they talking about? Are they talking about me? It ends up in this weird place where all the bots are just talking about us. But also at the end of the day, is that actually a bad thing? Because you know what sucks right now is when I have to email Steve and be like, hey, just emailing you a fourth time to get a copy of that bank statement. If you could get that on over here, uh, if you don't mind, kind of the last thing we're waiting on. Wouldn't it be nice if there was a helpful AI agent that immediately was like, oh yeah, Steve's already authorized me to say, share this stuff with you. Here it is. And boom, just like that, we've got that item back. I mean, there's a ton of like nuance in there, right? Like what about privacy and what's okay to share and what's not okay to share and who can it be shared with? All that stuff matters, but also isn't there a version of that that's better than the status quo, how it is right now? Especially when we think about how so much of our work is chasing people for documents. And there's a real risk when 80% of the emails that you send to a client are, hey, can you get me this thing? Can you get me that thing? Just following up on this to the point that when they get a notification from you, they see another email come in from you, they groan and they're like, oh, got to go hunt down some paperwork. That's the last thing that I want my clients thinking when they get an email from me. Like, they're charged— I'm charging them so much money. I want their perception of this to be like, oh, this is an opportunity. I get to talk with a smart person. Instead, right now it's like, here we go. Get the shoebox out, Tammy. Accountant just sent an email. Hopefully AI agents can make that better. On the subject of this, got a great question about deploying client-facing AI agents in an accounting firm. Get your pocket protectors out. We're heading over to the Nerd Corner. It's Nerd Corner. Nerd Corner. Let's nerd it out. Nerd Corner. Here's the question. I'm actively thinking about how my firm interacts with AI that clients are trying to use. I have a few clients really leaning into AI. They build out all sorts of custom GPTs and they want me to do the same. There's many considerations here and I'm not seeing a ton of conversation about it yet in the accounting industry. Mostly just people talking about how to make our own accounting work internally easier. I don't really want to like give away all my secret sauce, because that's what they pay me for now, is coming to me for that advice. But I'm entertaining building out a knowledge base of sorts and creating some, like, example projects with dummy datasets and then exposing some sort of AI clients could interact with that way. And our firm priority is fractional accounting, less tax, and very little just exclusively bookkeeping clients. Okay, let's talk about how you could actually do this today. Deploy a client-facing AI that they could come and talk to and get instant answers from. Because while it won't be as good as you, the real deal, an AI assistant can do many things that you cannot do. That is, uh, answer 24 hours a day immediately, respond to a question at 3:00 AM. Like, let's be real, you and I, we're super smart, but if AI can be an 80% version of you that is always on and always available, that will provide value to our clients. And the, the knee-jerk reaction is like, well, is that gonna cannibalize the value that I provide when they have to call me or something like that? I think like, to me it's, it's worth thinking about a bit, but the North Star has to ultimately be how do I like deploy value to as many people as possible in as scalable of a way as possible. That is what ultimately makes people happy to pay me. The alternative is how long can I white-knuckle this out doing it the old way before people figure out that, oh, the accountant across the street, they're pretty much like me, but that guy has an AI assistant. And so that's actually a more complete, better offering. I don't want to just like try to ride that one out, right? So if I were to do this today, here's how I would actually do it. Custom GPTs are not the move because they do not give you enough control over who can interact with them, unlike you on the backend being able to see what are the questions that are being asked and what advice is it giving. There are dedicated tools out there to do just this job. And right now we've talked about it before on the podcast. I would still send you to a tool called Delphi, D-E-L-P-H-I.ai. I'll put a link in the show notes if you can't remember those 6 letters, but it gives you much more control over like what questions will it answer and what quest— questions will it refuse to answer, being able to see logs of everything that people have asked it, being able to restrict who has access to it. Do they have to pay for access? Is it only available to people at certain emails? This is a fun experiment to go through if you haven't before. The idea of like, I started like, for me, this started as how could I build a clone of myself that would more or less advise how I do? And you're never going to get to 100% version of that. But how do you get to an 80% version of it? Because even 80% you is better than 0% you. And for that client who like needs a quick answer, it's going to be useful. So some things to consider if you are actually going to roll this out. First, what is ultimately the goal of the assistant? Are we trying to get them to something specifically like our services, or are we just trying to answer their questions? I would recommend you do the former because every time an AI assistant answers a question, while that person does build like a bit more trust with your brand, ultimately There has to be something, something in it for you in the end. So don't feel bad about like, hey, if you wanna get the response, like give me your email and I'll put it on my newsletter list. Or if they weighed into certain topics, like lead them to a landing page of yours. This is something that Delphi does really, really well. So what are these service offerings that are the entry points into your firm? Maybe it's cost segregations, maybe it's, you know, somebody got a very specific notice from a government agency or something like that. Delphi, you can set those things up so that anytime a conversation wades into that territory, it can take them on a more specific path to either gather specific information or hand them off to something, a product, a service, a landing page. So consider what are you trying to get them to, somebody that comes and uses this assistant, but also consider who do we want to have access? We could open this up to everybody. I've shared how, uh, Michael King, who is like an internet think person for, uh, fractional CFO firms, The link, the main link he uses on his LinkedIn is a Delphi AI assistant where you can come and pick his brain and it's, it's kind of backed by all of his thought leadership. So that's an example of an assistant that anybody can access, trying to go as wide as possible. But you could also build this so it's just for your clients if you wanted. Consider, are there versions of this that are different for specific client avatars in your firm? If I work with taxidermists and floral arrangements peoples, The AI assistant I would send each of them to is probably different because the more nuanced and specific the advice from the AI assistant, like exactly for taxidermists, exactly for floral arrangers, the better the assistant, especially when it has to be better than them just going to Claude or going to ChatGPT. Last, consider how will people find it? A lot of times you've, you may have done this as I have before, you sink a whole ton of time into making something. Something and then you make it and it's like your pièce de résistance and then nobody ever sees it. Nobody knows that it even exists. And you're like, oh, it's not actually about the thing. It's about the acquisition channel for people to find the thing. This is the biggest mistake that people make building a product or a course or something for the first time is they're like, I'm gonna spend so much time channeling everything I've ever learned into this, this monolith of a 12-hour course. When actually the better way to do that is to ship something fast that's that, you know, takes 30 minutes for somebody to go through because there's a whole lot more people that are willing to watch you for 30 minutes than are ready to commit to 12 hours. But then in the course of launching that thing, you learn like, oh, actually I have to be able to promote it and send people to it and figure out all these other things around it. So if you're going to have an AI assistant, consider before you go through the work of building it, how are people actually going to get to it? Will anybody actually get to it? Or is this just like a fun thing for you? And there's also merit in that, like just going through the exercise. Spring is here. It is spring cleaning season. The perfect time to clean out the dusty corners of your accounting firm and just get rid of the stuff that you don't want to be doing still. And for most of us, one of those cobweb-laden services that is just stuffed into the corner of the garage, it is bill pay. Well, good news, my friends. Artificial intelligence is here to help. You might have heard of it. The platform that's crushing it right now, this video's sponsor, Maker's Hub. Listen, nobody likes bill pay, okay? It sucks. But almost every business has to do it. And oftentimes it's high stakes. It's messy. It impacts inventory. You're going after like savings. The details matter. You miss a zero or a comma here or there. Oh my gosh. That's a big deal in bill pay, right? We are getting wildly good new AI models seemingly by the week right now. Right now. And honestly, MakersHub are the folks that I trust to like put this work— MakersHub have been like arguably making the best use of AI models over the last 3 years to solve a very, very specific problem. The problem of Bill Pay Nobody Likes It. So if you do this stuff for your clients, or you have clients who deal with project tracking, receiving, like construction people and manufacturing folk, HVACsmen, super recommend you check out Speaker's Hub, a great example of AI done well. To learn more, check out the link down in the show notes. Hey, just real quick, if you're using Outlook or Gmail to manage email inside of your accounting firm, brother, those were tools that were built, uh, 20+ years ago and not for accounting firms like yours. If you haven't heard of them, let me introduce you to one of today's sponsors, Missive. Missive, M-I-S-S-I-V-E. It is exactly what needed. I, I actually used a bunch of like off-the-shelf help desk sort of email tools in my accounting firm. Man, man, do I wish I had Missive. It's a vastly more efficient way to manage email across your entire team because email is not a hill that anybody should be dying on alone. We did a demo on the main channel of it. If you want to see it in action, just search my name and Missive. Really cool, doing some cool AI stuff as well. Worth checking out. To learn more about just what you're missing by using Outlook or Gmail, check out The missive info down in the show notes. Now, super important with your AI assistant is to really think through what are the sources of context? The only thing that's gonna make this better than just an off-the-shelf Claude is proprietary knowledge and proprietary perspective. How would you advise? And so we have a Delphi assistant. In fact, I'll put a link to it in the show notes. We'll occasionally use it to like go back and find like, you know, a 2-year-old reference that I made in a podcast. And it's, it's useful for that. But my proprietary context is 600 podcasts, several hundred YouTube videos. Like, and you're like, Jason, I got a job. I don't have 600 podcasts. So you gotta think about what does this look like for you? If you've ever done any writing, like blog posts, all that stuff is pretty obvious, but it can also be like offline stuff, stuff that you never publish, like voice notes. Like record a 10-minute voice note of talking through a very specific technical thing. Like transcript tools are so good now to where you can just chuck that audio file in, that effectively is your podcast. And Delphi, it will only ever give out advice that around things that you've actually talked about. Like it will say, I'm sorry, I don't have enough information to answer that question. And it'll even alert you to say, hey, people are asking questions about X, Y, or Z, and I need some more context here. And that can inform how you continue to improve upon the assistant. And so the next time somebody asks a question it can't answer, you can kind of treat that like Responding to a client question, like record a voice memo of you responding to that question. The next time somebody else has the same question, it can then answer it. How many times in the course of serving clients have we answered the exact same questions over and over and over again? This is an investment in an assistant to be able to answer those questions at any scale, 24 hours a day. We are scaling your essence, your je ne sais quoi, not as a replacement to you, but because even the 80% version of you deployed infinitely can do some good, can help some people. Now, importantly, you can't put client stuff into this. Like you can't put meeting transcripts into this. I'm frankly not even really a believer in redaction. I think redaction is actually very risky because you can miss stuff. There can even be hidden stuff that is not human viewable that can indicate a client that it's from or something like that. All of your client interactions should inform how you advise, but the advising that you put behind the bot, Make that advising in a vacuum, not like a conversation that you had with a client, but how you would hypothetically advise a client. I wish it was as simple as like just pumping in your meeting assistants. That being said, for an internal tool, something that you build inside of your Claude team or your ChatGPT team, that is definitely an opportunity to use your meeting transcripts, but that is more for your internal team to use, not for external clients. Now, opportunities to consider. What might make this worthwhile for you? Like things to be thinking about, ways we can use this bot. Consider what if someone discussed it with your bot would you want to be alerted about? Maybe it, it's ultimately like a potentially really high value client. Maybe it's a specific service offering where you're trying to do more of that. Maybe you want to specialize in beekeepers and you're like, oh my gosh, we just got our first beekeeping person talking with our assistant people. Everybody stop what you're doing. We got to get this person on the client list. It could be specific types of people, could be what they're talking about, but consider what you would want to be flagged about. If folks are having conversations with this thing. Also consider who else. If someone gets into like a specific discussion path, who else might want access to that person? Do you work with an investment advisor? Does your friend work with beekeepers? Does your friend run a, a just a payroll firm? If this chat assistant becomes a funnel where a lot of folks are discovering you, at least to some extent, there are more ways that that value can be distributed than just them coming to work for your firm. Last, and this is a big one, where else, assuming the bot works well, could this bot be deployed in useful ways? Could be an internal thing, could be used for training your team, drafting responses to client emails. I can tell you my Delphi assistant, which is like a snapshot of everything I've ever said on the podcast, which is kind of the novelty of this, like it's not making things up, it's parroting things that you have already said. That's a really important thing to understand. My Delphi assistant is accessible by my Claude. So let's say I'm having Claude kind of think through some ideas for, for a podcast outline or a newsletter write-up or something like that. If it ever has a question of how would Jason answer this, it literally has a bot it can immediately go to and get an immediate response of things I've actually said on that subject. Like that is a superpower. That is incredibly valuable for my AI agent to at any time be able to ask me a thought leadership question. And get an immediate response without me having to lift a finger. And again, like, this is an advantage of me having this body of podcasts, a huge body of context. And that's what most of us are blocked by right now is what is your body of context? But even your AI agent, like, think about it responding to an email for you, drafting an email. What if it had an outlet where it could go and ask this third-party service like Adelphi, uh, what would Jason do? What would the person say in this situation that ought to inform my response? The better that assistant becomes, the more like you it becomes. Truly, the more helpful you can be to your team, to your clients, and even to the AI agents that are trying to save you time. We're entering a really strange paradigm where all the software we've ever had to date has been around, like, automating what we do with data, getting a number from this system to that system, applying some sort of logic, making a decision of where it should go. But ultimately taking a number from A to B. Like, that is kind of what software has done for us to date is data automation. But we're entering now for the first time this era of knowledge automation. I can create an AI assistant that talks and acts like me. It's not perfect, but it's pretty darn close. What does it look like to have a fully scalable version of your specific expertise that can be deployed 24 hours a day at virtually no cost? That is knowledge automation in a way that we've never really been able to do before. And so how, like, you train your staff and how they make decisions, how clients interact with you in ways that may not even require you to lift a finger, or how you write your next social media post, like all of that stuff is going to be impacted by knowledge automation, but it will only ever be as good as the context you have captured. Does the bot know how you think? How complete of a model is that? Are we capturing it? Knowledge automation. I think it's going to get really wild. Last, we got some questions to get through. Let's go to the mailbag. Mailbag, mailbag. This segment is called Mailbag. Questions are submitted via email, jason@jasononferns.com, social media DMs, podcast comments, and even my telephone. 1-516-980-4968. Got a bunch of texts lately. Some of them I've even responded to. First question: I have a question on Double versus FinancialSense. I've watched all your videos on them and we're currently using Double. What is the difference between the two? Is one better for managing bookkeeping clients? We are in like peak tech selection season right now. Send me your this or that questions. This platform or that platform. I can give you the pros and cons of each. Because I've talked with tens, if not hundreds of users of each at this point. So Double versus Financial Sense. They both started out as platforms in fundamentally different places but have grown towards each other over time. So Double, previously Keeper, started as a month-end close automation platform. It connects to the accounting ledger of your clients and it gives you a single place to manage all of those month-end closes. Financial Sense started as a practice management system, which historically has meant task management, time and billing, and client portal. It's an oversimplification, but that's, that's kind of where that category started. Over time, people loved Double and they're like, we love that you can help us automate month-end close, but you don't have all the stuff that I need to manage my whole firm. And it would sure be nice if you were a little bit more like a practice management system. So Double shipped practice management features and now they're kind of sort of half practice management system too. FinancialCents, people were like, you're a great practice management system. I love you for this, but it would sure be nice if you actually helped me with more of my month-end close automation workflow. And so now Financial Sense is shipping more month-end close stuff. Now, Double does way more month-end close stuff than Financial Sense does. And I would say Financial Sense still has a more mature practice management offering than I think I would say that Double does. And so the right answer for which of these is the tool for you is a question of what do you want? If you're a solo person, if you run a small team, 5 people or less, those are the firms that in many cases are still getting by without a practice management system. But you get beyond like a single pizza— a single pizza-sized team, and you absolutely need a practice management system so everybody knows what they should be working on and where everything's at. There's a lot of folks that use Double also use a practice management system, but a lot of very small bookkeeping shops will exclusively use Double, and that's fine. So in your case, I would say it's a question of how big are you planning to build the team and do you actually need a PM or not? Next up, now that I'm past tax season, I'm going to send proposals to clients. Ooh, to get proactive with their tax planning. Good for you. Somebody listening to this is like, I'm literally golfing and haven't been back in the office since the deadline. Cool it, brother. Should this just be like a proposal for the tax planning engagement? Because I kind of want to use this proposal to engage them for the whole next 12 months. Yes, I am totally a fan of sending clients proposals for a 12-month agreement that are not at the end of the year. It's so weird how we arbitrarily stick to the 12 calendar months. I think there's an opportunity for your proposals and the timing of them to incentivize the behavior that you want from your clients. So if you want more of your clients doing tax planning with you in November, why are we sending them proposals in January or the week before Christmas? Whatever the cadences that you want to build your firm around, like when should the year start? If you're a tax shop, in my opinion, the year should start in the summer. Where you have your first touchpoint for like a 6-month check-in on how tax planning is going. You do that again at the year end, and then you file a return the next year, usually in the spring if you've been planning all year. And so a year's work starts in like May or June. That's when we should be sending them the proposal. And then by the following May or June, the tax return has probably been out the door. Like it's, it's filed now. But the previous year we were doing planning in the summer and the fall. And so, yes, absolutely. Don't just stick with sending clients annual renewals at the end of the calendar year. It's already such a clogged time of the year. Hard to get clients' attention, hard to find time for it around the holidays and all that. So absolutely, if you are feeling so inspired, that's the hardest thing is this time of the year, like having the motivation to like get your head back into the game pretty much. Obviously the US has this hugely if you do tax work, even bookkeepers, like they get into like March or April and they're just exhausted because you got the year-end crunch, you got 1099s, you got all the year-end payroll junk that comes up, and then you've got tax pros hounding you if you're in the US. Most countries have their own version of this, the crunch. And throughout the crunch, you stack up all these things where you're like, I'll get to that after the deadline, I'll get to that after the deadline. Then you get to the deadline and you're like, I am spent. And the last thing in the world that you then want is another crunch of all the things that you deferred and you promised during busy season. And it can— I mean, it can take weeks to get your head back into the game. That is what I mean. I had— I worked with a lot of people that always said, like, if only we could be as productive year-round as we are in March, we would— we would get so much more work out the door. And it's like, well, yeah, there's also like only so many marathons a human being can run in a row, right? But I'm a big fan of taking the annual renewal, sending the proposal to the client out of year-end because there's just so many other things going on at the end of the year. Next up, I had a client on my silver plan. So that is— that's referring to— they send, they send 3-tier proposals and broadly have a bronze, silver, and a gold plan. I had a client on my silver plan do an acquisition mid-year and scope grew significantly. What do you build into the engagement letter so that you can reprice without friction? I mean, this ought to be pretty clear for your client. Like, the disclaimer is nothing more than this is your price at the current level of scope. If your scope changes, we'll have a conversation. Now, baked into this question, they said my Silver plan, that's what they're on, but their scope just went up. I want to be really clear on this because this is the biggest misconception around how 3-tier proposals work and it costs people a lot of money. Buying a company when you're on the Silver plan doesn't mean that they ought to be going to the Gold plan. We need to disassociate Bronze, Silver, Gold, how premium an offering they picked We need to disassociate that from how much work we're doing for the client. You can have a client where you are doing a massive amount of work and they're still on the Bronze plan. You can have a client where you're doing a little itty bitty, a little tiny amount of work, but they're on the Gold plan. There are two dimensions to what you're doing for the client. There's the scope of what you're doing for them, and then there's the spectrum of how premium is the delivery of that service. And that is ultimately what the three-tier proposal is for. It is not for the client to decide the scope of work that you do for them, because I'm not going to let a client be like, well, I don't think we're going to file 1099s this year. Like, who's best equipped to decide the scope that they need? You are. You're the pro. So don't put a proposal into them that's asking them, hey, how much do you want me to do for you? You make the decision based on what I understand about your business. Here's what I'd recommend you have us do for you. And here are 3 different flavors of how premium that service delivery can be: bronze, silver, gold. You're selling coach, you're selling business class, you're selling first class. Because the reality is there's a lot of people that will pay 2.5x for first class, even though what you're selling is the same destination, just wrapped up in more access. They get your phone number, they talk to the big boss instead of the grunts. At 2.5x the price, that is a project that is by and large 4 to 5 times as profitable. So by giving them the option to opt into the more premium delivery, you could do one of those projects and dump 4 traditional, like bronze price package and dump like 4 traditional projects. You're doing a fifth of the work, getting paid the same amount. That is the power of the 3-tier proposal if you use it correctly as a framing for premiumness, not a framing for scope. Next up, thoughts on Perplexity's new personal computer launch. So Perplexity, this company, they, they were the first one to like nail AI search. And in the last couple of months, they launched a tool that like will analyze your QuickBooks file and be like, hey, I'm a, I'm a bookkeeper, I'm a controller. Let me look at your QuickBooks data and I'll tell you what you're doing right or wrong. And then they launch this thing that's like, I'm going to prepare your tax forms for you. And accountants are like, get off my lawn, man. All it is is a way for them to like attract certain types of people into their platform. More premium business users, get them on Perplexity. It's not like, well, we trained a proprietary model that can now do this or that. It really isn't. But this same company, they launched what they called Personal Computer. Which is an app that currently only works on Mac that is basically an AI agent that runs in the background. And it can use any old tool on your computer. And again, it does it in the background. So it doesn't get in the way of you using your own apps. It is just doing stuff for you. You can give it a task and it's humming along in the background and just reports back when it needs help or when it's done. The user interface, like it's exactly what we need. It's the sort of thing that should have been built into Windows by now. Microsoft has spectacularly blown the AI lead that they should have had. But we need the AI assistant that is working along in the background as we work. And you can just send it a quick note and be like, hey, go do this. Hey, go look up that. Hey, do this thing. Let me hand this app off to you and you can do that while I work on this other thing. Like that's the future is an AI agent that is just kind of working in tandem with you. That is sort of the interface like around how personal computer works. What I don't know, Does it actually work that well? You don't see as much about it in the wild as you do Claude Cowork, and it's asking a lot for you to hand your entire computer over to something that you don't have a lot of trust in. For the same reason that there's, I'm not gonna let like a guy walking by outside be like, hey, can you come? I've logged into my computer already. Can you come and do some stuff on here for me? That would be weird. This is kind of the software version of that. So I'm glad they're working on it. We'll see how it pans out. It's good to have more folks like pushing what does the user interface look like for our AI agents? Good to have folks working on that, but perplexity, like, I'm not jumping to it for now. Next up on the subject of AI models, here's the question. What I keep running into is firms will pick a Claude AI model, like, you know, Opus behind Claude or something like that. And then they give it all that client data and the model's really, really good, but the underlying data is a mess. And so the output just looks confident enough that nobody catches it in review. I think the conversation keeps going around like, what's the best model to use when the problems are actually upstream of choosing a model? Because there's problems in the data. This has kind of always been the problem when it comes to advising. Like, there have been so many startups that are like, oh, just connect us to your QuickBooks file and we will come up with incredible insights. Anyone who's ever known this, including the founders who've tried to build this stuff, the problem is always the quality of the bookkeeping data. And so that will still be an issue with any AI that is doing more like FP&A, sort of like big brain stuff on top of the accounting. But this is where, like, bookkeeping AI can help. And so there will be a— I mean, there's AI for, for everything. But yes, that doesn't change the underlying problem that, like, if the data going into it is problematic, then the output's not going to be any good. I would say that's, that's always been a problem. Hey, real quick, month-end close automation software. It has kind of completely changed what doing bookkeeping work looks like the last few years. And did you hear the news recently that this episode's sponsor, Canopy, they are working on their own spin on month-end close automation that is coming soon and will be built directly into the Canopy platform. Canopy, who you may have seen late last year, launched their own version of like tax intake built into the practice management system. Now they're going after month-end close automation, and it is all, it is all part of Canopy's vision to be the single sort of all-in-one place to not only manage the work of your accounting firm but to get that stuff done. I almost cursed. You know what I'm saying? Canopy always kind of been like a feature nerd's dream, especially if you are of the mind where you're like, man, I want to consolidate more of my stuff, eliminate some tools you're leaning maybe more all-in-one. Canopy has been one of my leaders in the practice management system category for as long as my recommendations have existed. Excited to see what they're doing in the bookkeeping automation space now as well. But to learn learn more about Canopy, check out the link down in the show notes. Hey, real quick, do you need to hire any humans inside of your accounting firm? I wish you could see what I'm seeing through my own two eyes right now. I'm staring at a portal of qualified pre-vetted candidates that are champing at the bit to come and work for your accounting firm. These are real human beings. Aubrey, a bookkeeper with 3 years experience, $3,100 a month. Samuel, an executive assistant with almost 10 years experience, $3,000 a month. Maycol, a bookkeeper with 6 years experience. It's got a CPA experience with, with Bill.com, Accounting CS, Deal, Power BI, QBO, NetSuite. Let's find a tax person. Mia, US tax accountant, experience with CCH Access, Drake, and UltraTax. Oh, bless her heart. $4,000 a month, gang. These are all super talented accountants that are ready to come and work with your firm through Cloud Accountant Staffing, longtime sponsor of the channel. They have a portal where you can go around and see all the folks apps that they have pre-vetted and even book an interview directly with those people from the portal. You don't even have to hop on a call with Cloud Accountant Staffing. All these applicants have— they've got resumes, they've got intro videos. Can't tell you how many firms that I know who have been very happy with Cloud Accountant Staffing. Accounting, it is a universal language. Luca Pacioli, buckle up for this one— not American, Italian. You know why? Because accounting knows no borders. Here accounting, there accounting, everywhere accounting. Accounting. Okay, you need some human talent in your accounting firm? Talk with Cloud Accountant Staffing. Check out the link down in the show notes. Next up, for cloud-based AI models, do you foresee individual companies continuing to develop their own models, or will AI become like operating systems are today? We essentially have like two operating systems, Windows and Mac, and everybody just builds on top of those. Is it even worth it for a company, say Juno or Soraban, to build their own model, or should they just link up with OpenAI or Anthropic? It's a great question. That ship has already sailed. Nobody is building their own generative AI models. OpenAI is closing in on being a trillion-dollar company. Anthropic is not far behind them. They have so many smart people. They have so many dollars to train generative AI models that are so unbelievably good that the notion of, you know, Juno, who just raised, I think it was $14 million. The notion of any companies in the accounting vertical going out and training their own model is an absolute fool's errand because it will never be as performant as even the generalist models. Now I'm talking about generative AI, which is one type of AI. If you look at a company like Digits, they use embeddings models to classify bookkeeping transactions. That is more machine learning. That is not generative AI. Those types of models, they can be built cheap to do specific jobs, but generative AI models like Nobody's building their own AI models right now. And so oftentimes in marketing, you'll see companies say like, oh, we train our own models, or it is a specialist model. It's not. It's a generalist model with some sort of wrapper around it. We have an AI assistant, a Jason bot that can see into all 600 episodes of the podcast. And when you ask it a question, it'll give you a pretty cringe response. It's like, hey buddy, yeah, let's see what I think about this or that, because it responds in the same tone of voice that I use on the podcast. That is not a tone of voice that anyone would ever write anything out in. It's kind of weird. But does that mean that I trained my own model? It doesn't. What I did was I sat down and I recorded 600 podcasts. And so we have all these transcripts and behind the scenes there's a generalist model that takes a question from the user and then it actually uses a different type of AI, an embeddings model to go out and look across all the transcripts and say, what are the times in all these transcripts when Jason Staats like talked about something that was the most similar to that. And then it pulls all that stuff in, the actual words that I said, and it goes back to the generalist generative AI model and says, hey, here's a bunch of stuff that Jason said before. Here's the question. Make an answer for this question using more or less the words that Jason used. Now there's a, there's a bunch of nuance in like, how are those transcripts chunked and how big are the pieces and how many pieces do you pull in? And does it actually use my own words or is it allowed to make its own words? There's a bunch of nuance there. But at no point did I train my own generative AI model. It's a generalist model that is using stuff that I said to give a response. And so to the question, this ship has already sailed. Everybody's using OpenAI and Anthropic. I mean, look into it. Every month you get a press release. Intuit's partnered with Anthropic. Intuit's partnered with OpenAI, aka we just committed X millions of dollars of spending, basically the enterprise version of switching from monthly to annual. And then they write a press release about it. Hey Jason, was wondering if you had an opinion on a CMA cert. Oh, my least favorite class. I took auditing my sophomore year and I thought there's nothing worse than this. And then I took management accounting. I know you specialize with CPAs mostly. You take that back. I don't specialize with CPAs mostly. And I saw your video about how becoming a CPA doesn't really do anything. Okay. Just makes you one of 600,000 CPAs in the US. Are you just trying to trigger? All the CPAs that listen to this, the main point is the learning, not the certification. I'm thinking CMA will teach me a bunch to be like a CFO and do planning for small business. Okay. A lot of things to unpack in that question. What I've said about the CPA is that the CPA doesn't achieve for people what, what undergrad education tells you that it would achieve, which is a $10,000 bonus from whatever Big Four employer that you're working for. I guess it does achieve that if you go to Big Four, but also it doesn't give you like this level of scarcity that I think it is sort of built up to give you because there's over 600,000 active CPAs in the US. And so if you get your CPA and you go out and you run a firm, the next thing you have to do is somehow become one of one, like somehow become an expert and become the go-to person for that thing. Whether you have a CPA or not, that is the exact same. Like you don't get the CPA and then not have to do that. If you have a CPA, you still have to become specialized because there's literally over 600,000 active CPAs in the United States. So the CPA is useful for some things, but not for becoming like this hot commodity, you know. Now the CMA, you will for sure learn some stuff, but I would largely say the same thing about CMA. Like a knock on the CPA is it is a big bar. It is, you know, you got the whole 150-hour thing. It's a very expensive, very time-consuming bar where there's some people who are like, if I don't get my CPA, I can't run an accounting firm. I can't be an accountant anymore. And it's like, whoa, whoa, whoa. That is absolutely not the case. Many, if not the majority of the best accountants I ever worked with were completely uncredentialed. And I worked at completely reputable accounting firms. I see what that sounds like now. Like I was working out of a hybrid, like testosterone shop, massage parlor, but we'll also do your tax return in the back. You could absolutely get a CMA. There will be some people that are impressed by it. But on the other side of getting the CMA, you still have to do the hard work of what am I known for? What's my reputation like? What is my specific type of expertise, I don't think in this day and age anyone's going to see your CMA and be like, ooh, you must be a real smarty pants, as much as you've been on 6 podcasts for taxidermists and you're the taxidermist CFO now. Like, that is the path to being perceived as the expert in a thing, which grinds our gears because we're like, well, what does that have to do with technical expertise and actually knowing your stuff? That is the unfortunate nature of professional services firms is perception is reality. We benchmark ourselves on how good we are at the technical stuff alongside other technical people. The customer benchmarks you on— was that— could I see his unmade bed in the background of that video? Nice guy, but poof, might have had tuna for lunch. Like when they don't understand the technical stuff that we do, they will make buying decisions based on the part of the process they do understand, which means you gotta make your bed. Gotta sort out that mustard stain on your shirt. I mean, you gotta think about how the customer perceives the whole experience of working with you in ways that have nothing to do with the technical stuff. So my answer for the CMA would probably be more or less my answer for the CPA. Last one. We recently acquired a firm and are newer to marketing. Is in-person networking still worth it post-COVID? Boy, it's been a while since I've heard the phrase post-COVID. And what do you recommend most between social media content, BNI, and strategic referral partnerships? I would say first things first, it depends on whether you are specializing in local or specializing in like industry niche. I would say US tax firms are probably the only remaining type of accounting firm who can really make a strong argument for specializing in local because state and local tax in the US is so hard. Serving an industry niche, if I'm going to just work with beekeepers all over the country but also do their tax work, very, very hard to run an efficient tax business. Serving clients all around the country. That being said, a lot of people do it. So most firms are either going down the path of specializing in local or specializing in industry. And it's not to say there's not a ton of other dimensions to specialization. That is kind of the first generalization you can make. If you're specializing in local, most social media platforms are not very good at local. The best one is Meta platforms— Facebook, Instagram. They actually do local pretty good, but otherwise the in-person stuff is very, very important. Especially with local. If you are specializing in a niche that is nationwide or even global, online stuff, really good. Ultimately the highest leverage way to find people. But still, at the end of the day, nothing beats in-person. So if you're running a local specialization where you're showing up at a lot of local events, you should probably still be doing the same volume of in-person events. For a national niche, that in-person stuff is just happening all over the country now. That in-person, I mean, just cuts through so much noise in a way that online stuff cannot. But the holy trinity is the combination of online and in-person. That was only two things. The holy binity. Those two things working together, and I've seen this in my own events, we can get people to come to in-person events because of the online stuff. And so like we have this cool unlock of in-person being able to be cool in a way that it couldn't be had it not been for all the online stuff. And so the answer is both of those things. BNI in this day and age, there's people that have got good stuff from BNI. That's not where I would start. Always look for the choke points for a specific type of person, the bottlenecks where they all have to come through. If there's some sort of required licensure to get into what they're doing, what does that choke point look like? Like, I mean, the Holy Grail is the example we've gone through of an accounting firm who works with a specific type of a specific grocery franchise. And everyone that's going to own one of those grocery stores has to go through this accounting and tax training program. And they fly this accounting firm owner out twice a year to do the training, and they get 80% of the people that are going through that pipeline. That is the dream accounting firm, right? And some industries have those choke points. Some don't. A lot of them, you maybe just haven't found it yet. Like dental, dental supply providers, like dental supply folks, they're sending sales reps in person to these dental clinics. They are regional. So let's say you're going to try to build a local dental niche, which is tricky. You better become best friends with that person that is showing up in person and has a relationship with all of those dental clinics already. Like these, these bottlenecks exist and they're always the highest ROI place to find clients. Now, if you are at, say, a different stage in running your firm and throw out that if you're at a different stage now, one other kind of overlooked channel here is getting your work from other firms. It is so powerful. I'm pretty much— every accounting firm has clients they're ready to move on and would do so quicker if they had a great referral partner. And a great way to do that, buddy, coming to my live in-person workshops. We almost made it to the end without me plugging it. I'm coming to 20 cities in the land of the free. In fact, we ran the numbers on this. There is an event of mine in the next 2 months within a 2-hour drive of 52% of the US adult population. Okay, 52%. Take a look to your left. Take a look to your right. One of those people is within a 2-hour drive of coming to my event. But if we expand that to 4 hours, 4 hours, I mean, that's— we're still talking about a one-day thing. 4 hours. It'll take me 2 hours to do this podcast. Okay, meet me halfway. The percentage of people in the US who are within 4 hours of one of my 20 events, 85%. Look to your left. Look to your right. There's a 50/50 chance that all 3 of you are within 4 hours of driving to one of my events. If you're coming from a ways away, see if you got any other locals you could carpool with too. Like, it was very normal at every event. We had people that were 4, 5, 6 hours driving out to these events. But like, doing an event in Houston— in Austin, loads of people coming from Houston, loads of people coming from Dallas. See if you can hook up with some locals if you are making that trek. The workshops start in the early afternoon for this reason, because a lot of people drive for them. It's also why the dinner at the end of the night is optional. So there's a cheaper ticket if you don't want the dinner, because I know a lot of people drive out for these events and that'll give you time to drive back without throwing away the money you spent on a dinner ticket. Closing in on 1,000 signups, gang. Gonna be a ton of fun. I'm gonna be on the road basically for 2 months, but it is fantastic meeting a bunch of people in person, but also seeing all the connections that you make, which from the last time we did the tour last fall was unbelievable. The stories that have come out of that from referral relationships to acquisitions to everything you can imagine. It's just like, it's wild getting 50 to 100 people that listen to the podcast in one room, like, who are all like local people. Like, everybody's leaving with new friends. Like, the vibes are just incredible. Want to see you there. JasononFirms.com/events to grab tickets. Gang, thanks for popping by today. There's a lot of podcasts you could have listened to today, but you chose this one. And to me, that means the world. Whether you're a 4-monitor guy, doesn't matter to me. Okay. Just glad you made it here today. Any questions, pop that stuff down in the comments and I will catch you Catch you in the next one.