616 Your Accounting Firm Pricing Is Broken (Here’s how to fix it)

Listen up, did you hear that Jason on Ferns is going on tour? Buckle up, it's the show you love, now with snacks and an open bar. The tour's halfway through, but we still need you. For tickets and more, head to Jason Staats. Jason on Firms.com. Jason on Firms.com. Now it's time we end the song. On to the show. Today on the podcast, one of the most common things I hear from people, like where their accounting firm is just stuck and the thing they can't get over, it is pricing. We talk about it so much, but today we're digging into the 6 most common pricing traps and how to overcome. Each and every one of them. Also, AI nerds, you might have heard new model in Claude. I'll be totally honest, having a full existential crisis right now with how good Fable is, the new model. Got some things I want to talk about there. Also digging into the mailbag. Where is it okay to use AI in my accounting firm today and where should we not let it touch? That was the wrong way to say that. How do I vet the people that I'm bringing in to my accounting firm to know if they can really hack it? What to do if your clients are concerned about your accounting firm using AI and 2FA codes, bank login codes. Oh my gosh. If you have to log into different systems for your clients to be able to gather these things, usually in the context of month-end bookkeeping, it's a nightmare. Still a nightmare. Okay. We're going to talk about how to mitigate just a little bit of that nightmare today. Friends, let's do it. Welcome along, my friends. How do you do, fellow accountants? My name is Jason Staats. Welcome to the Jason on Firms podcast where we talk all things accounting firms. I'm sorry for the transition you just saw of an AI version of me line dancing, but we are closing in on the end of the Dirty Dozen, the 12 biggest, most common pains that are ailing accounting firms today, as sourced from over 1,000 real live accountants that came out to my in-person events last year. Today, pricing. You sick of talking about pricing yet? Might be time to rename the podcast. The Pricing in AI Podcast. Pricing, obviously it matters. There's literally no more important thing than pricing in your accounting firm. It is what will set ultimately the profitability of your work. We know this, right? But it is the solution to overwork, to having enough money to hire the right people to support you. It all boils down to pricing, not workflow. Oh, do we love to talk about workflow and geek on systems and all that? All of that is secondary to what price can we get people to pay for us to do our work? But it's which just comes with a lot more emotional toil, a lot more tough feelings than getting excited about that next AI thing, right? Let me run you through the 6 most common pricing problems that I see in the wild. And these are becoming an even bigger deal now with AI sort of heating up the conversation around how our work is perceived, the value of it, but also how that work gets done. And so the first problem here that we really need to be on top of is a lot of firms are still stuck on hourly. All of their work is going out the door according to how many hours went into that work. And you know, at this point, inputs do not equal outputs. Like an hour is not a good yardstick for how valuable is the thing that I just handed off to the client. And our clients, frankly, could care less how much time we spent on a project. They have their own perception of what that work is worth. Now, for a long time, This has been the easiest proxy we can use to price our work. And at midsize firms and beyond, honestly, it is still basically the only way to pull 1,000, 10,000, 20,000 accountants into a system and have a way to consistently price people across all those human beings. Like, there is a reason that every big firm works off of hourly pricing. But you know what? This podcast? It's not for your big, big 4 firms. Okay. I know y'all are listening. Keep listening and you will go and start your own firm. But that is the beauty of running a smaller firm. When I say smaller firm, like let's say 100 people and under. The beauty of running a smaller firm is you have more control to do stuff like this. Get off of hourly pricing, get to project-based pricing. So easiest way to step into this as someone myself who rolled this out in a 40-person firm on, uh, across over 2,000 projects a year. This was an 80-year-old firm. It came with all of the blockers, every, all the counterarguments, everything that you would expect, all the reasons for why this could never work. Here is the easiest way to roll this out. Anybody can do this. You start with recurring work. People can fight until the end of time on whether or not you should, you should price hourly for one-off projects. I don't really care about that. You would start with the easiest stuff to move to upfront pricing here. And that is your recurring work, once a year work, if you're doing bookkeeping, monthly work. Step 1 is to give every client an upfront fixed price. And this is not a promise that the project will absolutely be this price at the end. What this is, is the price they will pay assuming scope is more or less what it was last time. That is all that we are communicating. You can even take this as far as giving them a range, but that introduces some complexity that I think is avoidable because ultimately, What we have to look at is how do we protect this part of the business, the profitability of this part of the business, not every single project. There is so much time and toil that will go into making sure that every single project is priced as fairly and accurately as possible, as opposed to, hey, we do 600 of these. How do we make sure that we just make a little bit more money across the 600 and not get so worried about like getting into the weeds on every single project and on how we made out. This is maybe the hardest thing for traditional hourly folks to divorce themselves of because they go to the people who bring you the mess and it takes a bunch of time. And then you go to the other people who bring you that stuff in an organized way and it doesn't take that much time. I do still think there's a world where you could price those two projects differently. All we're doing now is spitballing that upfront. And the fallback then becomes, you can set a boundary ahead of time to say, if the actual time that goes into this falls you know, X percent outside of what we expected, then we'll have a conversation with the client about rescoping the work and a price change. And all of this is explained when we give them an upfront price. This is what we think it will be. And by signing this engagement letter, you're actually giving us permission to pull funds to charge you this as soon as the project is done, completely eliminating accounts receivable and having to chase people down for payment. In my firm, in our case, we just set a boundary for WIP at the end of the project. Basically, if the project came in under 120% of what was expected, it would get auto-billed out. If it was more than 20% over what was expected, it would go to the account manager to review it and see if we needed to change the price. That very first year that we did this, I think it was 94% of projects went out auto-billed within that WIP range. That was 94% of projects that somebody wasn't having to sit down and manually bill. It was a huge win. And honestly, clients appreciated it. Like nobody likes the ta-da moment of seeing the bill from your attorney or from your accountant after the fact. And you just feel like you have no control over that and no idea what it's going to be. They appreciate having that price up front and it makes your life easier too. It goes, it should go without saying with AI stuff happening right now, and we're going to talk about the new model from Claude, our inputs have never been less relevant than they are now to our outputs. You can see one firm, that is like really leaning into pulling Claude into production and getting the most out of this stuff and getting incredibly efficient because there's very real situations right now where you can trade human hours for machine hours. And I always come back to the example of, let's say a client brings you this mountain of receipts and you point Claude at a folder of those receipts and have Claude rename every single one in your preferred naming convention, like date, payee, amount. And it goes through and it does 700 of these receipts for you. And now you don't have to drill into every single one, at least not to the level that you had to before and have someone sort of manually rename them all. If that saved you an hour of time, are you still going to charge the client as if a human did that hour of work? The answer is yes. And there are so many other examples of this in our firm right now where real human hours can be traded out for machine hours. And at least right now, we can still price as we've always priced. Who knows if 3 years down the road that's still the case? But today, that doesn't make your work any less valuable. But what it will create complications with is traditional hourly billing. So it's never been more important than it is right now to get that stuff figured out. Pricing problem number 2, what should I even charge? This whole like shooting from the hip, pricing by gut, pricing by what you think that thing is worth. Which honestly doesn't really matter. What matters is what the other person thinks that it's worth. It's not you paying the bill, it's them paying the bill. And oftentimes people are like, well, yeah, but that pricing isn't fair. Like they're agreeing to pay a price. If they don't think that it's fair, they won't pay it. And the only perspective that matters on this pricing is like, what is their perspective? If I'm saving them $1 million and then the next person, they're like, I just, I'm paying you to do a tax return because I just, it has to be done. Those two people will not value you the same way. And so you can absolutely price those projects a different way. That's not contingent fees. That's not unfair. If it's unfair, they'll say no. These are adults. But when we start setting the price based on like kind of how we feel about that work, we end up frustrated when they don't perceive the value of what we do the same way as we do. And this is all of the science of like positioning your accounting firm and finding a rare and valuable pain that you can solve for people where they can't compare you to the accounting firm across the street and be like, well, that guy does it for $200. Why are you charging me $1,200? The opportunity here becomes— this is where you get into value pricing, like pricing work based on the perceived value of your work. Now, I didn't do this on 100% of the projects in our firm. We ended up doing stuff for people that we didn't really want to do and didn't want to put that much effort into, like 1099s as an example. In the US, 1099 information reporting, is this super irritating thing you have to do in January. All your clients want you to do it, and I just want to spend the least amount of time on this possible. Sidebar: AI is going to completely change this 1099 process this coming January. But I would do upfront value pricing for kind of our core work, the, our bread and butter where we actually made money. But stuff like 1099 work, like I just made a turnkey pricing, it was like the first X X number of 1099s were this flat amount, and then for every 50 in addition to that, it was some other amount. And we just tried to basically make that as turnkey and quick to machine through as possible while not sinking so low as to end up in these projects where you have to do all this work for the clients only to find out you don't have to file any 1099s and now the client doesn't want to pay anything for it because you're not actually issuing any forms. And this would become this one-off negotiation that everybody in the firm would do with every single client. In cases like this, it made sense to just draw a line and say, here is the firm policy, here is the price for the first 15 1099s, here's the price for each sort of additional block of 1099s. As we went more premium there, we also developed some really good free resources to help clients DIY the small projects that we didn't want to have to touch ourselves. The better your free DIY resources for your clients, the more confidence you will have with your minimum just going higher and higher on work like this. But I'm an advocate of value pricing your core work. That's how you drive really good margins. But then the other stuff, like, make it as simple and turnkey as we can make it. Third pricing problem, the underpriced legacy client. Your longstanding clients being priced way below market. And another sneaky version of this is your low-bono clients. The clients where you're happy to give a bit of a discount, but before you know it, they're referring all their friends who also expect a bit of a discount, and now a third of your of your client list is these people that are paying like 20%, 30% below market. Now the worst version of this is the people who were with you from the early days and they're still paying the price they were paying in the early days and you have this loyalty to them. You feel like you can't cut them, but at the same time you have this resentment of them where it's like, well, I'm not excited about the work I'm doing for them anymore. I'm excited about this thing over here, but they're still on the bus and not paying me very much to be on the bus. This is like all wrapped up in like, you know, not getting too big for your britches and forgetting where you were and all that. We got to be real in these situations. Oftentimes the reason these people are still with us is because they can't get a better deal anywhere else and they are not getting your best. A really easy trap to fall into is to think that nobody will help these people as well as we can help them. Isn't it funny like how we look at every other accounting firm, someone that's that's like a firm we pick up a client from or a firm that a client leaves us to go to. We look at these firms and they're like, "Poh, they just don't get it that we are so much better than all of them." Let's be real here, people. There are other firms out there that can help these clients. And if they're a client that you're no longer excited about, they are almost certainly not getting your best and would be better served by somebody else. This is made way easier by finding a good referral partner that you trust for whom they are a great client Like so much of this is as simple as do they have your attention or not? Are you trying hard on that engagement or not? Like, this is really 80% of the game. And for someone that's at a different stage, for whom that's a great client, they will have their attention and they will do a much better job than you will in all likelihood. This is a great case for like showing up at local events like we host to meet a bunch of other local accounting firms that are sort of in different stages. As soon as you have a referral network, it gets so much easier to send these clients along faster which makes room on your plate for the stuff you're really excited about. And I get that it's hard. I, at one point, cut a third of our client base. This was in an 1,800-client firm. I totally get that that's hard, but what's the alternative? The alternative is also very hard and gets even more existential because you have to deal with the reality of like, "Oh, I'm committed to all these people, but they're not getting my best, and now I have to carry the weight of that." That's not the way to do it. Pricing headache number 4, is just the raw fear when it comes to raising prices. We are not, I don't know that accountants, most of us have the tools to deal with these emotional things. And it's why we geek out on workflow things and tell me about the next AI thing and all that. But having hard conversations, doing things that you know people won't like, that's tough. And it's like an identity thing. Like all these people that we meet, these folks that we support, you want them to think, kindly of you, and we worry when we increase prices, their perception of you, we're afraid of what happens to that, right? But one of the biggest lessons that I learned in running a firm and having to fire a bunch of clients and have a load of those phone calls was this gets so much easier when you separate the personal from the business. And starting out, I think most of us, it's all personal because we are living by like every single engagement, trying to put money on the table. And so when somebody comes in and buys into your firm and will pay you fees for doing your work, that feels like an incredibly personal thing because you are so appreciative of that. And they are the ones that are helping you kind of go out on your own for the first time. But the risk here is if that whole personal side is never sort of separated from the business side, then almost inevitably when your business is going in a different direction than their businesses, The divorce becomes a very personal thing, right? That is why in most cases, firing feels like this big, scary thing where you're like, what if I bump into them at the grocery store? Like it's this big sort of dramatic thing. It's the parents of some friends of my kids. I couldn't do that. That is a sign that the personal was way too wound up in the business. You can maintain a personal relationship with somebody while your businesses are going in different directions. If you decide that you're going to niche into beekeeping, and going forward, we're only taking in beekeeping clients. All that takes is a conversation with your garage door installer client that says like, hey man, like we found the thing that we're really excited about and it's taking us in this direction. Obviously you don't fit within that, but let me help you find someone who's going to take great care of you in the next phase. And I'd love to stay in touch and I'll see you at the, I don't know, at the end of your slip and slide party. This is another argument for annual renewals that are like, here's the contract for the next 12 months and establish this and setting the expectation of, hey, we're going to revisit this every single year to make sure that it still makes sense for everyone. A lot of times we shy away from the fear or the work that goes into doing these annual renewals and we're like, no, we just have a standing thing that auto-renews so we don't have to have that conversation. But that, that whole expectation of, this is a forever marriage is what makes the divorce so scary. And we can't hide from that. Like, we need to learn to have confidence and be able to be frank in those conversations. Otherwise, you end up stuck in a bunch of engagements that you don't actually want. As I read this ad, I am feasting my eyes upon talented accountants that you could literally, literally book an interview with right now. Like Ruru, a CPA, a tax pro, $3,500 a month. He's got experience with CCH Access. Bless your heart. CCH Engagement, QuickBooks Online. Gang, he is one of many accounting pros that you can peruse at the Cloud Accountant Staffing portal. We'll link it down in the show notes. It is a veritable smorgasbord of accounting talent, folks that they have pre-vetted, where you can go in, you can see, see their experience, you can literally see an intro video that they've recorded, and you can book an interview with them without ever even having to hop on a call with Cloud Accountant Staffing. Amazing. We got tax pros, bookkeepers, auditors, executive assistants. Oh my gosh, I got a lot of people in here right now. 45 candidates. Because I know you need that next dream hire. You really do. What do you got to lose? Okay, go check out the portal, book an interview. Feels like a win-win, right? You're only going to move forward if you find somebody you love. Cloud Accountant Staffing, run by folks that used to run an accounting firm. They make it as easy as possible for you. To learn more, check out the link down in the show notes. Hey, if you're done licking your wounds from tax season, if you run more than a few hundred 1040s a year, you gotta look at this episode sponsor, Sorban. One of my leaders in tax workflow automation can save you a tremendous amount of time right now, specifically on the intake part of 1040 workflows, but they're also working on connecting to your tax software automatically, like pushing data in there. They're working on automating the delivery of the tax return, but this is like the fastest growing category in tax tech. If you're a US tax firm, you absolutely need to have a good tax workflow solution nailed down. We did about 1,800 1040s a year in my accounting firm. This would have saved us so much chasing clients for documents and just tedious organizing of all these PDFs. Oh my gosh, can't say enough about Sortman. We actually did a full demo of their platform on my YouTube channel not so long ago if you want to see it. But to learn more about Sortman, check out the link down in the show notes. Two more to go here, and these are around the biggest traps with three-tier proposals, which I love that a lot of accounting firms have now gotten into because this is how you really optimize the profitability of an engagement as a 3-tier proposal. Trap number 5 here is the 3-tier bronze trap. You launched your bronze, silver, gold packages. Everyone's still picking bronze or even worse, they're just ghosting you. They're not responding to those proposals that you send out. And let's, let's cover that first. Why would people be ghosting your 3-tier proposals? Most commonly, it's one of two things or a combination of those two things. One, They are overwhelmed. They are confused. You send one proposal to a client that is just this huge dictionary of all these technical things that they don't understand, they're overwhelmed from one. But putting three of those in front of people is like, "I have got no freaking clue, man. I don't know what to pick here. You tell me, you're the expert." And so this is the most common problem with accountants doing three-tier proposals is they put way too much into them, too much technical stuff that the other person doesn't understand. The other reason why they ghost is we don't put a deadline on it, is we just send this over with some vague expectation that they will do it at some point rather than there being some urgency built into that process. Now, more generally, if everybody's opting into the bronze tier, this ought to go without saying, but it's because they're not seeing anything that's exciting to them on those silver and gold tiers. And this is usually because the way us technical people, accountants, the way that we set up these three-tier proposals, is we load them with technical things that might be interesting to us, but the client probably doesn't even understand what those things are, or they've never had those things before, and so they don't understand what the value of them would be. The things that the clients actually want usually are more access to you rather than your team, maybe a faster response time, and it's very simple things around like, what does it look like to simply get help? And can my team come to you directly even rather than me having to be the one to interface with you? It is not this list of technical things. And so make sure that those tiers are built around what people actually want. Right now, honestly, is a really good time to be doing annual renewals and get it away from the calendar cutoff. You'll, you will be very glad that you did in December. But those renewals really need to be more built around what's the behavior that we're trying to drive. What are the things we want them opting into? And if you run a tax firm, that is planning. You want them opting into tax planning. You don't want them just coming back to you when a return has to be filed in tax season. You don't need more of those clients. You need more planning clients. And now mid-year is usually when you're doing that first planning check-in, and that is the best time to kick off a renewal that runs through the completion of the tax return next year. On the bookkeeping side, I like having a standing like mid-year project. The end of the year is a lot. It's a mess. And if you're in the US, you're doing 1099s at the end of the year as well. I like to have a mid-year project that can be something like a systems analysis where you are looking at like, what do their financial ops look like? What are the apps that they're using? Is there anything that we need to change? Let's at least have this conversation once a year. But this can create a great sort of kickoff point for renewals mid-year as well to get that stuff off of the end of the year. Everything happens at the end of the year. And before you know it, you're getting into busy season and you're like, I don't have time to do all of this stuff because I was just— it was the holidays and it all happened so fast. And now I'm into February and I just like, I just, I got to do work. There isn't time for the admin. Last pricing trap is the 3-tier gold trap. When people first roll out 3-tier proposals, I would say 2/3 of people struggle with the previous issue, everybody picking bronze or being ghosted.. But then there's another third where all their clients pick gold and they're like, now what do I do? Because usually the first time you roll out the bronze, silver, gold thing, on that gold plan, you are giving away the farm. You're giving away everything you've ever done for people. Like, like you will watch their kids on the weekend. Like anything you can imagine, you have now promised on the gold plan, but everybody just said yes to it. And you might be getting paid stupid money where you're like, I can't believe people are paying us this much. But you're also like, I have no idea how I'm gonna get all this work done now. Simple solution to this, just from a strategy standpoint, when you're sending out proposals, don't send them all out at once. Send them out in batches. This could be 5 batches. You know, think about it as sending it out in 20% blocks. Start with the people that you most want to keep, who you think are most likely to say yes to the most premium stuff. Send those proposals out. Don't be afraid to go premium on the pricing. Obviously, if everybody's saying gold, there's probably a pricing issue here too, or we haven't reached far enough with gold. My take here is bronze should always be a bump over whatever we did for them last year. Silver should increase by the profitability of bronze. So if you're a 30% margin firm, a 40% margin firm, the bump to silver should be 30%, 40%. And if you're a solo firm, you got to factor your time and your costs into this, the kind of the overall profitability of the firm. But then gold ought to be a big step up. So if you made a 40% jump from bronze to silver, I'm gonna make like a 120 to 160% jump to gold. We don't want very many people saying yes to gold. But in practice, to mitigate against the risk of too many people picking gold, you gotta send these proposals out in batches. Say 20% at a time. If you sent out the first 20% and everybody said yes to gold, then that's a good thing that we just sent 20% of them, right? Pricing in that next batch going way up. And people struggle with this. They're like, "Well, that's not fair. The people in the second batch are paying more than the people in the first batch just 'cause they were in the second batch?" Yes. There is no such thing as a fair or unfair price. If people are saying yes, then they are happy with the price. And if they're not, they won't say yes to it. But through this sequence of doing 20% at a time, this is how you dial it in to take on the gold volume that you actually want, which should not be a very high percentage of your projects. We're down to the last few sort of big meta issues, the dirty dozen facing accounting firms. But if you think there's one we missed, a pain for your firm, put that stuff down in the comments. We'd love to see it. Maybe we can dig into that in a future episode. But it is time to get your thick rim glasses out and head over to Nerd Corner because we have got a wild new AI model to talk about. It's Nerd Corner. Nerd Corner. Let's nerd it out. Nerd Corner. So a couple months back, Anthropic, the company behind Claude, came out and said, we have this new model called Mythos and it's so good that we won't even let you mere mortals have access to it. It was supposedly a model so strong that it would bring about all sorts of cybersecurity issues. And people were like, eh, I don't know if I actually believe you. But just this week, we got a model from this Mythos family called Fable. And right now, if you run to your computer and you open up your Claude, you will see in the model selector that you have an option for Fable. And you can think about this as a new, even more capable model above Opus, stronger than Opus. You might have to install an update on your desktop Cloud 2, by the way. When you open it up on the bottom left, it should give you a notification that there's a new version. Update to that and you will see Fable. But this is not just another incremental step. This is a big step. I'll be honest, like the last 12 months, really until Claude gave us some amazing new models last December, I was kind of bummed at the quality of models getting better. Like it was, it was almost like smartphone launches where it's like, okay, like it's a little bit better, but it's not that big of a deal. Fable 5, this new model, it's a big step. And the things people are sharing it's capable of online, like it is mind-blowing. And this has a very real impact on how we plan our work in an accounting firm because there's already been this trend toward like, okay, maybe this thing isn't an AI assistant as much as I need it to actually like take the lead on production, like the actual doing of the work. We put out a bunch of videos around like how this is very real now, like Claude is good enough to do the heavy lifting. And with this new launch of Fable, I think it makes this all even more important. And it requires rethinking some of our systems to actually be more AI agent friendly. Because if we're putting together work right now in a way where AI really can't be helpful to it, we kind of gotta stop and be like, okay, what, what's a version of our system that we still like that is much more friendly to an AI to be able to come in and, and help progress the project. A good example, I'm at AICPA Engage in Vegas a couple of days ago, and I'm talking with a guy that does audit. And in audit, most of your work papers are done in Excel. And for a while now, Claude has been phenomenal at getting stuff done in Excel, like so, so good. And you can absolutely imagine a world where if 80% of work paper production in an audit is happening in Excel, like goodness gracious, can a lot of this stuff now be driven by Claude. Unfortunately, when it comes to tax work, when it comes to month-end closes, there's more stuff happening across sort of different systems. But to give you an example, in tax, a lot of firms don't have like a sort of core workpaper template that they use. They just take the source documents and they put that stuff straight into the tax software. While there's other firms that have like an Excel workbook that they use as kind of a tie-out. So they actually enter the stuff from source forms into this Excel workbook. And then also enter it into the tax software. People that don't do that are like, why are you doing that? That's a waste of time. It's an extra step. People that do do that are like, well, this is how we like tie out the totals on the tax return and make sure that everything is correct. I would say the right answer here for what should be done has completely changed now that we have Claude, which can absolutely build this Excel workbook, complete it for you like that work now. I mean, by and large can be completely automated by Claude pulling information off of source docs, even out of like emails and putting it into your Excel workbook template. Getting that data into the tax software, that is still a bigger lift, although I would argue it is something that now is easier to push down to an admin level than it's ever been before. Especially this is another benefit of having that workbook template is if you've templatized how all the information is pulled out of the source documents in your communications with clients, then it actually makes it much easier to standardize how that data from the workbook is then entered into the tax software. And you're not quite as reliant upon your big, expensive senior people doing all that work themselves. And so that's, that's a very practical example of how changing a system from one version to something that maybe didn't make sense to you before AI will actually yield a much more productive process where you could have AI doing the vast majority of that workbook preparation, if not all of it. It then goes to an admin for entry into the tax software, and you are now 80% to project completion by the time it hits the desk of your professional staff. Obviously, there's a lot of nuance in there that we just cruised past, right? But our systems as they are, as they've been built for humans, may not necessarily lend themselves to an AI agent be able to come in and be helpful. But as they get better, these agents, as the models get better, it becomes more and more imperative that we have a system that we can actually pull AI into. Now, Fable, incredibly capable, but as you might imagine, there's a dark side to this, and that dark side is the cost. In fact, it is so costly that as I record this, Anthropic has said that Fable's only going to be made available for the next 2 weeks. Like they can only guarantee you're actually going to be able to play with it for the next 2 weeks. They're talking about potentially any usage of these models being pay-to-play. And so as part of your subscription, they're saying you might not actually get to use these models at all as part of your ongoing subscription. And if you do want to use them, you have to pay for the usage. And people, uh, the, the cost, it is not cheap. In fact, for the first time ever, an AI model is being paid above minimum wage. Sort of not a joke. If you standardize it, if you assume that an AI model is generating 40 tokens per second. And obviously this is a, this is a huge assumption because different models generate stuff at different speed. But it assumes like if you're doing an hour of work, like one person might be back and forth doing a bunch of stuff with the AI. Another person may not be doing that much with it. But this is just kind of a fun way to like benchmark the cost of an AI model. Assuming it's generating 20 or 40 tokens per second for an hour, the previous best model, which was very expensive, by AI model standards, Opus 4.8, that model would have been making between $14 and $18 an hour. Compare that to Google's arguably their best model right now, Gemini 3.5 Flash, it's making about $8 an hour. GPT-5.5 was between, exclude the GPT-5.5. Fable-5 at 40 tokens per second, you'd be paying at $40 to $43 per hour. That, I mean, that is like, that is not to say if I'm using this, then I'm actually paying that, forking that out per hour. But you've seen more and more news stories lately of these big enterprise orgs who will like cut headcount and say, and do these really aggressive AI rollout strategies where there's an expectation of the team to consume a certain number of tokens. Like they have to use AI for things. It's not optional anymore. And then the AI bill goes through the roof and is more than the cost of all the humans that they just laid off. And then they have this, the CEO is panicking and they're like, oh my gosh, you have to pay for AI? I think we're entering an era in the context of small accounting firms where we will be paying attention to cost on this more than ever before and trying to optimize the subscription and the usage that's allowed for within your subscription because you're already paying for that. And that's actually the cheapest model usage you will ever get and how to get the most out of that. Right now, your model limits reset every 5 hours. But if you're like me, you sleep 5 hours a night, which means if you have Claude at the limit and you're hitting your limits throughout the day, you can schedule Big Lift stuff to run overnight and, and consume some of that usage during non-peak hours. If you're on a team of 10 people that are, that are using Claude, every member of your team can schedule stuff to run overnight so it's not consuming your usage during the day. I think we're still in the early days of that becoming a real problem for people. Fable is actually making that a lot worse because it consumes your usage pretty fast. We're in this 2-week window where they have doubled Cowork usage limits. And I, I think it's so people that don't just immediately run out of usage because of Fable. But it'll be interesting to see first what Anthropic does in a couple of weeks if they take Fable away, if it becomes exclusively pay-to-play, but also this whole conversation around usage limits and what people are actually willing to pay for their AI tools because we're talking about a world where AI is actually doing production in a way that our humans have historically. And so you could see being okay with paying more than we've paid for software in the past, but not everybody's there. Like, I mean, I think the most hard done by people with the current pricing model for AI, it's still solo firms where they're like, well, I want to use the business plan because from a security standpoint, arguably you have to, but the minimum number of seats you can get from Claude on their team plan right now is 5 seats. So am I really going to pay $150 a month to only use one of those seats? Right now you're kind of left with no other option. But you can also make the argument that like, am I going to get $150 of value out of this thing? Almost certainly if I'm pulling it into production work. So this whole like the math around like how much AI are we going to be using going forward, it's getting really interesting. Model optimization will also get more important. The idea that you use the big expensive models for planning the process, but then it delegates down to cheaper models for more of the grunt work. I mean, does this sound like anything you've heard before? Like, I don't know, maybe an org chart where you're like, don't use Fable for data entry. Oh gosh, use, use our admin staff Sonnet. It's so much cheaper. Those conversations are going to be getting much more real once we're banging up against our usage limits. How do we effectively sort of delegate tasks to cheaper models where it makes sense? Hey, I know you love Cloud Cowork as much as I do, but have you heard the news about Canopy Coworker? Subtle. It's a new AI assistant they launched inside of Canopy, one of this episode's sponsors. Canopy, a practice management system that is a feature nerd's dream. They're trying to cover as much of everything you need in an accounting firm in a single place.. But as we talk about on this podcast often, it is much easier to roll out an AI agent inside of a closed system than it is to roll out an AI agent that works across like all of my systems, right? That's the trade-off. I want the AI that works across all my systems, but we're struggling now with like, what's the most secure way to do that? The AI agent that works inside my all-in-one system, buddy, that is easy. Flip the switch, count me in, strap in, buckle up. I want it. Canopy is so convinced that you're going to love what they got to show you, they're going to cut you a check for 75 big ones. They'll give you a $75 gift card. Gift card to Bed Bath and Beyond. What's that? Okay, I'm not sure where the gift card is for, but if you book a demo with Canopy via getcanopy.com/jason, I'll put that link down in the show notes, they give you a $75 gift card just for coming and demoing. Do it. What do you got to lose? Are you still running your accounting firm on a spreadsheet? Puh. Don't do that, man. That's gross. Financial Sense, it is the practice management system for small and growing teams. Think 20 employees and under. It is everything you need to get your work done all in one place, and they've been shipping some cool stuff lately like built-in month-end close automation, this deep integration into the accounting ledger to actually pull all the transactions in platform and do that sort of quality assurance pass from inside of Financial Sense. Now they also recently launched a campaign made called Proudly Small, because being small isn't something to apologize for. It's something to be proud of. And just to be clear, we are talking about your accounting firm. Wouldn't it be nice to have a tool that's actually built for small firms? That's actually a very common complaint right now is tools are increasing in cost. They won't come down to support firms beyond like a certain number of clients. Financial Sense, they are built for you. Plus, If you use code JASON10, it will make the price of Financial Sense even smaller for your small firm. You'll get 10% off the annual team or scale plans with 2 or more users. I'll put the link to claim that down in the show notes. Check that out. See, Financial Sense might be for you. What is accounting coming to? You know what? Cloud accounting stuff came around and all of a sudden accountants are like, what do you expect us to be like, technologists helping small businesses get set up in the cloud for the first time? Like, I'm an accountant. Then COVID comes around and everybody's like, who's going to help everybody with this COVID relief stuff? And they're all like accountants. And you're like, buddy, I've already got a full-time job and I'm trying not to get sick. And then AI comes around and it's like, what does it look like to run an accounting firm now? Well, now you're going to need a PhD in knowing how to do AI stuff. Welcome to 2026. I know you're overwhelmed. Me too. We all are. Everybody take the hand of the person next to you. Let's be overwhelmed together. Tell you what, it's great for me. We will keep putting out AI use case videos. I don't know. I don't feel like I've really done a good job of like building a business around the whole AI thing right now in a way that many people are. There's so many people that are building like AI education businesses that have no audience and put out no information for free. Yet here we are. I've got a full-time team of like 15 people and we are just pumping all this free content into the ether. And everybody's like, hey, can you come in and like train our firm? And like, hey, can you come do this? Come do that. We're like, no, no, no, no, no, because then we won't have time to make all the free content. I don't know. It's interesting. When I sold my firm back to my partner, ChatGPT came out 2 months before I did that. Of all the times for like me to go all in on thought leadership on accounting firms and a lot of tech stuff, like of all the times for that to have happened, it is wild. I am thankful for it, but it's a lot. Speaking of a lot, I just took a look at the mailbag. We gotta knock some of this down. Okay, let's get through some listener questions. Mailbag! Mailbag! Mailbag! Mailbag! This segment is called Mailbag! Questions are sourced from YouTube comments, from social media DMs, from emails I get, from texts I get. My number is 1516-980-4968. Shoot me a text, leave me a voicemail. First question, how can I vet whether a new bookkeeper that I'm going to hire can handle the workload of one of our experienced bookkeepers before I even hire them? So how do you know that somebody can hack it when you've never even worked with them before? It's always amazed me how much we will just hire an accountant and assume that they can do a job based on vibes without ever even having seen them prepare a tax return or do a month-end close. I've really pushed people for a long time to do trial days. This is the biggest unlock that I had in my hiring process is doing half-day trials with people to just get a better sense of what are they actually capable of from like the actual production work down to simple things like when the trial day starts, a member of the team will hop on a screen share with them and we'll send them a zipped folder. And we want to see what happens when they're faced with a zipped folder. We want to see how they use a computer for 5 minutes. You can learn a lot about how someone works by watching them use a computer for 5 minutes. Imagine how many people you have hired without even having this very base level of knowledge about how this person works. But this question is more to the end of, if my experienced bookkeepers can handle 50 projects, how can I know if this person can also handle 50 projects? The short answer is here is like, You can't. Like, there is no having perfect information to this. But how do we get closer to that? I would say, one, in your trial day, you will get a sense of how long it takes them to do stuff. We will have— I mean, the trial day saved me from making so many bad hires. And you also have to give people a lot of grace in the context of a trial day. They have technical issues. If the stress is too much for them, like you do have to be very reasonable and not expect too much from folks in this sort of controlled environment. But you will see one person take 5 times longer than another person to do the same thing in a trial day. That is great intel to have up front. So you want to know as much as you can about how long does it just take them to do stuff. You also want to look at their work history. This is something that you can suss out if you sort of follow the breadcrumb trail of where have they last been. I had one person that worked for me that was tremendous and they went to work for a firm that ultimately like downsized the team. The firm went from like 8 people to 4 people. And this happened 6 or 9 months after they hired this individual. Individual, and they were not one of the ones to get cut when they went down from 8 to 4. And I saw that and I'm like, I mean, that's kind of all I need to see is like they must have come in and done a great job improving themselves very quickly. That's a really positive signal. Last is to get an understanding if this person is, is a firefighter or not. And in my firm, I had people who needed stability and that was their happy place, kind of the stability, the repetitive stuff. But I had other people who were lit up by change, by urgency. I had a member of my team for whom the tax deadline was their favorite day of the year. And a lot of people, they hear that and they're like, "That's nuts. I can't imagine that." You need a mix of those people on the team because if everyone on your team is overwhelmed by change and by urgent things, then when that stuff inevitably happens, it wears everybody out. They all hate it. But if you have the person on the bus that loves being the firefighter, that becomes their job, the triage. And they can protect the people for whom they more prefer the stability and they're able to do their thing. The firefighters are able to do their thing. This is something you need to know about somebody and it's something that they can usually communicate to you pretty clearly. Like, do you, do you feel like you're more the type of person that's like, I want to sort of have my monthly system for how I do these things? Or are you the type of person that really enjoys change and special projects and the urgent thing coming up and you're like, yeah, this kind of fires me up. You want both, both types of people on your team because the firefighters will also get bored by the monotony of the month over month. So there's not a right answer for what you need. Just make sure that you got a balance. Jason Staats did a quick search and couldn't find you touching on 2FA and passkeys anywhere. I'm talking about the 2FA that's required for logging into client bank accounts when you're using their credentials and a code is sent through to their cell phone or the cell phone of a member of your team. As our virtual team grows, I don't want these codes coming to staff cells to protect the boundaries of staff and client security. It goes on into passkeys as well. Passkeys are kind of a more advanced problem. I would say the more realistic thing that most people are dealing with right now is how do I get access to my client stuff when the only option is client— is logging in as the client and then having to do two-factor authentication Olympics. First things first, our staff should not have access to client credentials. They should not be able to see the login, the password under no circumstances. Should that be the system? And if you're just using a basic password manager right now, your team can see the credentials. This is a massive problem because if somebody leaves your team tomorrow and they had access to 50 different client credentials, now what? Like, do you have to reset all of those? Most of us don't. But like, obviously that's a very real problem. The only solution to that is a single sign-on solution, something like an Okta, Okta. We had Practice Protect in the accounting space, which is really the only SSO solution for accounting firms. They were acquired by RightWorks, and I believe it's now called Cloud Protect. And I think you can buy it on its own without any of the rest of the, the RightWorks stuff. It is a pain, these systems and setting up all of these logins in a single sign-on system. But what this looks like in the end is your team, they log in to a single system and from there can click through to access any of those client logins. Not the credentials themselves, but like they will click Wells Fargo, John's Car Wash, and it will take them to wellsfargo.com and automatically put the credentials in and log them in without your team being able to see the credentials. So that's a very important thing when it comes to— it is a bigger thing with bookkeeping practices than with tax practices. When it comes to protecting client credentials, I don't rely on traditional password managers. You need a single sign-on solution. When it comes to two-factor authentication, I will give you my good, better, best, which is a terrible framing because good is not good. We should not call it this. Let's call it bad, better, best. Bad is us using our clients' logins, sharing a login with the client, the code going to the client. You've all, we all get these codes. You get a text message. In fact, let me pull one up on my phone here. Here's one from my personal bank. First words, all caps, don't share this, period, ever, exclamation mark, maps CU code. 741927. It continues. We will never ask for this. If someone is asking, hang up immediately. How do you think your client feels when you're like, okay, this is gonna come in in just a minute. Yeah, now can you send me that? Can you tell me what that was? No, don't worry about it. This is just our system. Don't worry, you can trust me. That's, gang, that is just not the way to manage accessing these systems these days. So that is the bad way to do it. The better way to do it is to centralize where these codes go to. And this will work with some systems and not with other systems. The way I did this was literally a physical phone through Verizon was $20 a month. And anytime a message landed on this phone, it was forwarded to a Slack channel that my team had access to. There are other systems like Twilio and like soft phones, uh, that will give you a phone number. But many of these two-factor authentication systems don't work with those sorts of phone numbers. The only complete solution I found for this was literally an actual phone. That I got through Verizon. And then all of those codes were forwarded to Slack or Microsoft Teams or somewhere where the rest of the team could see it. Now, not every like bank login where you have to share a client login will actually support this, like another device that you can send those codes to. But it is still better than the code going to the client. The best solution here, though, like the only complete solution is just getting onto better tooling that supports the actual secure way to manage this stuff. And so if we're talking about banking, You look at like Mercury, who's— they've sponsored content in the past, but like I— all my banking stuff is on Mercury. It was before they started sponsoring stuff. Mercury, if you're an accounting firm, your client's Mercury account can be connected with your accounting firm and then you as an admin in the accounting firm can give access to members of your team and you can dial in at a very granular level what level of access every member of your team should have there. So you're not using the client's login. We not only have our own login, we have firm-level access that can then be shared with members of your team only to the level that they need, like read-only, even like down to the bank account. And you can force two-factor authentication for any way that your team members access this account. I mean, it's like the way we access any normal SaaS app these days, but done for banking. You know, if you use like Gusto or something like that, It is how a business would give your firm access or QuickBooks. They would give your firm access to the QuickBooks file and then you would assign the people in your firm and their levels of access to that specific company. There are banks that do this, spend management platforms like Ramp that do a really good job of this. If your client is struggling with this issue, don't do what I did for a really long time and apologize for the inconvenience of them having to send us the code and all that. Be very clear. The system you use right now, like, is out of date and does not facilitate secure ways to access your stuff. And we are happy to help you move to a better system. Don't be afraid to have that conversation because it is the only complete solution to what we're describing here. You do have to take this security stuff seriously. I had a 6-figure-plus insurance claim in my accounting firm when a client's IT environment was compromised and then tricked our team into doing something. And this would not have happened had we stuck to the tools that were supposed to be used in that specific process. Basically, it looked like the owner was trying to bypass that stuff to fast-track something, which was not the first time that they'd done that. And the person that got access to their email literally copied the thing that they had done in the past. So it sounded just like something the owner would actually ask you to do. But we gotta take this stuff seriously. And if our clients are not on the right systems, we really need to encourage them to get to something better so it's easier for us to support them securely. Next up, due to the environmental, economic, and financial impacts of AI, how do we address clients and peers who are opposed to that AI? There's a lot of great reasons to not like AI right now, but the funny thing is, and we've, I'll start sounding like a broken record before long 'cause we've talked about this on the podcast quite a bit, but there are certain things that everybody can agree that, oh yeah, AI should definitely help with this. I don't think anybody would be upset if AI cured cancer. They would be like, What a great use of AI. But there's other things that we almost all universally hate, like AI taking the job of an artist or looking at AI art, or if the next podcast episode I published was just an AI version of me, we'd be like, that's lame. Like, I'm not interested in that. And when it comes to like AI coming into your job, all of us are like, I don't want AI to come and replace what me and my friends do. And there's situations where human beings will unite and be like, yeah, don't use AI for that cool creative thing. But the reality is, is what we do, I don't think is one of those things people will rally around. I think AI being able to prepare somebody's tax return feels a little closer to AI curing cancer than it does AI taking away some cool human creative thing. Most people, they despise what we do and that's why they pay us happily to do it. And so I think realistically, Like concerns over an accounting firm using an AI to get their work done, that's probably a very short-term thing. Most of these people would probably be thrilled if they could just rock up to an AI that would just do this stuff and they didn't even have to work with us. And at the simple level, like the bottom of the market, that will be the case. That stuff will happen more and more. I mean, you take like, let's say you had niched into creatives or voice actors or translators. I wanna talk about a, a, category that's just been absolutely destroyed by AI. Translators. Let's say that was your niche before. The people you work with, translators, hate AI for good reason. It has displaced them, their friends. Nobody likes it. But how durable of a, of a business is it for you to say, yeah, we're an accounting firm that will never use AI because of just how much it's hurt our clients? How long will those people happily pay $1,200 instead of $300 for a tax return because they're going to work with an accounting firm that doesn't use AI. Even those people, I have a feeling, will come around to being happy to use AI to prepare their tax return, to do their bookkeeping. That just doesn't feel like a sustainable, like, long-term place to set up camp. AI hasn't quite replaced accountants yet, but if there is a part of accounting that is in an endangered state, It is accounts payable. It is the AP department and the company that I think is taking AI the farthest right now, specifically inside of AP. It is Maker's Hub. They have taken automated extraction from invoices and receipts farther than I've ever seen anybody take it, but have followed that through all the way to like complex approval processes. You know, like when you need a delineator that is more than is bill over 'x' dollars or under 'y' dollars. When there's a little more complexity to it than that, MakerSub's got you covered. They handle payments as well. They're doing foreign exchange stuff. They are doing Canada stuff. It is an elegant solution to a problem that I think we all wish didn't exist in the first place. When you have ugly bill pay needs, be it from project tracking, inventory management, whatever that is, and the normal tools, they're just not cutting it, not getting the job done. MakerSub worth checking out. Putting AI to work on the most hairy accounts payable workflows. Learn more about MakerSub. Check out the link down in the show notes. 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 ever more green 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 going to see every single little thing that I put out. Goodness sakes. There's just too much. Mm. If you're not on my newsletter yet, check that out. JasononFirms.com. I'll put a link to it down in the show notes. Next up. I sold my CPA firm in 2019 and was never going to do accounting again. I'm now in the process of getting my CPE up to date and my permit to practice in public again. I'm bored and too young to retire. There's so much noise going on with AI now. That I'm thinking before I begin marketing my firm again, that I need to learn about AI before I do anything else. I'm 61. Amazing. But I've always liked learning new software and new shiny objects. I started following you on LinkedIn because I'd seen some of your YouTube videos. What do you think I should focus on next? Amazing. 61 years old. I don't know what I'll be doing when I'm 61 years old. I'll probably be working. Everybody needs to feel like a sense of purpose and feeling like they're helping people, right? That is why I, even if I could stop working today, like I wouldn't because like my work is, like I feel like we're helping people and everybody needs some version of that. Not that it necessarily needs to be work, but yes, lots going on with AI right now. But I will say if you're feeling behind, there's a lot of AI nerds right now that are very on top of AI that don't know how to build a business. They don't have a network, don't have the connections, don't know how to actually like effectively price on an accounting firm and do all of these things that a lot of the more seasoned folks that don't have the AI expertise have much more experience with. So don't feel like you have missed the boat because maybe you're not up to speed with AI. I'm going to let you in on a secret. Pretty much everybody feels like they're behind on AI right now because I think we probably all are. But yes, learn AI and build your firm. I wouldn't just learn AI because you will have to keep learning AI. Like it's not as if You can go deep for 3 weeks right now and be like, okay, I figured out AI. Like, it will have to be an ongoing thing. But yes, do both by all means. Getting back into the game. Love to see it. Is it possible to use Claude to work inside of QuickBooks Online to do repetitive tasks? Simple to explain stuff like delete a bunch of duplicate register entries. Yes, you can. So this, the feature here is Claude in Chrome where Claude can take over a Chrome web browser on your computer. The trouble is this is kind of the riskiest way to use an AI agent right now is on the open web. You can tell it, do this work in QuickBooks, never leave QuickBooks. You're only allowed to use QuickBooks. And it is exceptionally unlikely that it then wanders into another website. The risk of it just banging around on the web is it can be tricked. This is prompt injection. Instructions can be hidden on a website to tell it to do something that it shouldn't do, and it might be tricked by it. The Claude models, like there's all this red teaming around this that prevents it from following malicious instructions. A really annoying thing right now is it will oftentimes stumble into things that it thinks are a malicious instruction and it's like, no, that was a legit thing. Just do the thing. And so it's exceptionally unlikely that your agent is going to be tricked by something. But we have to be smart about this stuff. And for me right now, I draw the line at just not letting my agent use the web. And that's a bummer because that, that does eliminate some really cool use cases. But even one step shy of that, not letting your agent use the web, there are so many cool things we can still do without it banging around in Chrome. I would say a more like a, if you want a more secure way to automate stuff in QBO, consider building a Chrome extension. I mean, check out Write Tool, which is a Chrome extension for QuickBooks and lets you automate a whole bunch of stuff and it's free. They have a free ad-supported version, but Claude can also build you a Chrome extension. Like you can use Claude to build a Chrome extension that just lives on your computer. And this is arguably a safer and maybe more responsible way to automate things in the browser. Now the limitations will be, um, like you have to basically hard code what are those automated things that it can do. And you could click the Chrome extension and there could be like a dropdown menu of these different options. That's one way you could automate things inside of QBO. Uh, but to answer the question, yes, Claude can do stuff inside of QBO, but that is the one aspect of Claude that I'm honestly recommend disabling right now. In fact, that's the next question is related. If you disable Claude in Chrome, do you have another way to have Claude view and evaluate webpages as you work? Um, you can always send Claude a screenshot. If it's a public webpage, you can still give Claude that URL and it can go and see everything on a webpage. But I imagine you're like filling out a form or something like that. No, the unfortunate answer is if you disable Claude in Chrome, Claude cannot hop into a browser and do things for you. It's funny, this is where I've drawn the line on sort of AI security stuff is I think in Claude, in the organizational settings, you need to disable Claude in Chrome because while it is cool that Claude can go out and do stuff on the web, this is the riskiest thing we can have Claude do for us right now. And there's, it's still such a big spectrum of people who are like, AI's not secure, you can't do this stuff. And I'm like, yeah, it is if you have the settings right. And this would be my conservative positioning in my firm right now is disabling Claude in Chrome. But then you have other people that are like, "Ah man, but it could do so many cool things if we would just give it access to this." Like that's never enough, you know? I wish I didn't have to be a security expert. To be clear, I'm not a security expert, but it's really frustrating that the most common question that I get right now is like, "What's the secure way to use AI?" Like why are there not security experts that are more forward with this information, not paywalling it, like just putting more helpful stuff out there? Again, I would say, see how far you can get with building your own Chrome extension. Chrome extensions are really cool for automating stuff on webpages that you regularly use, and Claude can absolutely build you a Chrome extension. And that Chrome extension for you, totally secure to use. It's just on your machine, just in your browser. I mean, if you use Write Tool, it does amazing things. Imagine a base, a very basic version of that that you could set up to work across all the different apps that you use. That's really cool. Next up, when you upload a bank statement to AI, are you redacting account numbers? I can give you my answer to this. No, I'm not a believer in redaction. It is true that redacting identifying information from documents makes that non-sensitive by just about any rules. HIPAA rules, GDPR, California's privacy rules. If it's been redacted of identifying information, it's generally not considered PII or sensitive information. That's why Google can sell all of your, that's why all of your medical data can be sold to Google. Because it's been anonymized. But in the context of running an accounting firm, my policy cannot be with my team. It's okay to do this, but you have to make sure to redact anything that you put into the AI. That's just not realistic in practice. Like you're going to miss something like that. That is not something to rely on. So in this case, either you're okay putting bank statements into the AI or you're not, and approach this in as first principles and educated of a way as possible. Understand how, how Claude works, how large language models work. We've got a ton of content on this stuff. There's a lot of people that have absolutely no problems doing this. As I've shared, I have no problems doing this in my firm as long as we have Claude and Chrome disabled. But don't rely on redaction. Redaction is like, this thing isn't secure, but we'll still use it with sort of only redacted information. I don't really want to tiptoe into that. That's not realistic when you have a whole bunch of members of your team and you're expecting them to redact. Things. One other thought on Claude in Chrome, because there are a lot of like real AI nerds that are like, yeah, but I still want to have a play in Claude in Chrome. You may be much more equipped to do so responsibly than your team. And so, and that's how I am. Like in my case, if I were using Claude in Chrome, I actually have plenty of confidence that I could direct it in a way that won't be problematic. The problem is the whole rest of the team. In the Claude organizational settings, you can only disable or enable Claude in Chrome for everybody.. And so if the only way you're going to be able to use Claude in Chrome is enabling it for all, I don't recommend that. Maybe down the road we get more granularity here where you can enable it for specific team members. That would be cool. This question is on the video we put out recently of using Claude to replace QuickBooks using an open source tool called Bean Count. Really, really cool. If you haven't seen that video on YouTube yet, go check it out. Question is related to this. I started building a Claude and Bean Count setup a few weeks ago when Jason Staats mentioned using it. It still doesn't feel ready for primetime though. I don't like all the data being local to me. How can I pass on the, the accounting history to anyone outside my firm? Once all this same functionality can be done in the cloud, then I think QBO is toast. But right now it's still like the CD player that you plug in to the tape player in your car that doesn't have a CD player yet. Okay. You're upset that BeanCount gives you a local backup of an accounting file. Brother, put it in a shared drive. Send somebody else the local copy of this file. Am I old? I mean, this is what QuickBooks Desktop was. We bounced files around. It was no big deal. It is true. This isn't very suited for like collaborative accounting. If you have a whole bunch of people working on an accounting file, this isn't the move. Like, this is for the very, very simple accounting files. But for me to rock up to Claude and do some accounting and have a single file that then has captured all of that accounting, and I can share that with somebody. I can save it somewhere, come back to it later. I got no problem with that, man. As soon as it goes into the cloud, that is one more software subscription for you. And there's situations where I'm happy to pay that ransom, but many situations where I have all these holding companies and rental LLCs, all these simple little businesses that my clients set up, I don't want yet another software subscription for every single one of those companies. Now, to your point of once this is in the cloud, I think QuickBooks Online is toast. I mean, the closest example we have to this right now, Kik, is like an AI accounting ledger. They've launched an MCP, which means your cloud can connect to Kik, but it is a read-write MCP. So it can not only pull information from the accounting ledger, but it can push and write to the accounting ledger. So it can book transactions. You can go to cloud with a pile of 12 bank statements and say, go do the accounting in Qik. And it can import all the transactions into Qik and can classify them. That is the future, is something like a Bean Count, but is like this cloud-hosted, like very light accounting ledger. I don't know what the, like the market and the pricing for that will look like because there's this really cheap, like sort of utility sort of version of this where people just aren't willing to pay very much for it. Nothing close to what like is commonplace for Xero and QuickBooks Online right now. If your cloud is doing all the accounting and you just have this backend that is hosted in the cloud, people are not going to value that backend much. And you can make an argument, you can roll your own backend for pennies on the dollar and just park the bean count text file somewhere. And that costs you, I mean, literally next to nothing, pennies. But yes, agreed that as soon as someone gives you a persistent way to like store your accounting ledger, the accounting database, and just do the work through Claude. You still need a reporting layer there as well. But you better believe, like, QuickBooks is shaking in their boots. Like, at least for simple files, a full-blown SaaS app starts feeling really like overkill when Claude can do that work, right? Two more to go here. Does a standardized chart of accounts across clients still matter in an age of AI? That's interesting. Standardized chart of accounts. So historically, where we will standardize our chart of accounts across clients is if you have a company group and you're doing rollups like consolidated reporting, then you really want the chart of accounts to be the same. The other scenario where you do this a lot is around benchmarking. And so if I work with 100 beekeeping clients and one of the cool incentives of working with my accounting firm is we will like do your financial reporting, but you'll also see some benchmarking data around kind of where you sit and anything that is maybe out of whack with the rest of the market. In that situation, the more similar the chart of accounts is, the better. Now, does this change at all with AI? I will say the handling of different chart of accounts has gotten easier with AI. A lot of people have shared they're doing like 10+ company consolidations with Claude where you just, you throw all the financials in, you say, produce some consolidated reporting for me, do some eliminating journal entries. Claude can do all that stuff really, really well right now. Even if the chart of accounts are not the same, Claude will figure out the groupings and pull that stuff in in a logical way. Pretty much the same way you would expect a seasoned accountant to do this. The reality is, though, is if two companies are not using the same chart of accounts, the lines are drawn slightly differently around what bucket an expense gets put into. And so if your consolidated reporting is very general and there's like 5 buckets you care about, probably not a huge deal for the chart of accounts to match perfectly. But the more granular you want that reporting to be, the more it does still matter. How similar those chart of accounts are. So it depends on how you want to use it. In some ways, maybe it's gotten a little less important in an age of AI, but it all depends on like how decisions are made and what information you need for that decision making. Last one. What's your take on the Microsoft Copilot of today versus 6 months ago? The cloud integration with Copilot Cowork, pardon, oh my gosh, seems to be a way to get some of those tools in more account hands, presumably because it's part of the Microsoft ecosystem. If firms are not willing to commit to standalone Claude, I see what you're doing. You're trying to get me to give you permission to use Microsoft Copilot because they now use Anthropic's models. But using Anthropic's models does not make it Claude. This is very, very important to understand. This is like saying one accounting firm is the same as the other accounting firm because they both use QuickBooks. No, no, no, no, no, no, no, no. Microsoft can use Anthropic's models, but what makes Claude really, really good— I mean, they have good models, but what makes it amazing is the harness. That is the system that Anthropic uses, Claude uses for deploying these models and assigning little sub jobs out to sub agents. And all this orchestration happens anytime you ask it to do something, then it comes back with this amazing novel result. Claude is good as much as because of as much of the models as it is the harness. Heck of a sentence that just was. And it's why Microsoft, 2 months after ChatGPT came out, they took a 50% ownership stake in OpenAI and they said, we're going to put OpenAI's models in our Copilot so that it's— Copilot's really, really good and is as good as ChatGPT. And then years and years and years went by and Copilot was nowhere near as good as ChatGPT. This is no different than that. The only thing that's changed is Microsoft has now acknowledged like everybody else that Anthropic has the best AI models, not OpenAI. But that doesn't mean Copilot is going to be worth using for the same reason that for a long time ChatGPT was the clear option over Copilot. Now, has Copilot gotten better in the last 6 months? Yes. All AI has gotten unbelievably better in the last 6 months. But I would still make the argument that your time is too valuable to be using anything but the best. And Claude is light years ahead of anything else when it comes to doing actual production at work. It doesn't mean that there's not this inferior AI that can help you write an Excel formula and all that stuff. That's great. But when you're talking about a profession where people bill $1 to $500 per hour for their work and we're not willing to pay another $30 a month for a Claude subscription, I mean, this is most common in big firms where they won't greenlight it from a security standpoint, or somebody's bought in, like drinking the Microsoft Kool-Aid too much and are probably removed from the reality of working on it and maybe not realizing just how vastly inferior it is to the best tools right now. But I don't think your time should be spent on anything but the best. We actually had somebody at one of my live workshops that raised this question because they worked at a firm that was pretty old school and wouldn't touch anything but Copilot. Come to find out, this person had come to the firm or had come to my workshop without their firm's knowledge. And by the end of the night, had found another accounting firm to work for, which can definitely happen at my workshop events. They meet a whole bunch of progressive people and they're like, holy geez, what am I doing still working here? If you're one of those accountants, by the way, where you're in a firm where it feels like you are rolling a boulder up a hill every day and you're like, I see it. Why does nobody else see it? Look into my eyes. There's a better firm out there for you. There really is. And I don't care if they've promised you the moon. Anybody with a pulse and an IQ above 35 has been promised the partner track. In 2026, that is the sad reality. But if you feel like you're the one pushing a boulder up the hill every day, the answer is almost certainly to find someone else that you are better aligned with. This is the entire reason why I bought a staffing company a couple of weeks ago is I have so many firms that are listening to the same podcast that you're listening to right now that are just like clamoring to find that next really good hire, somebody that's aligned with them and their vision for how to run a firm. There are so many of those firms right now that are hungry for people like you. Stop applying your efforts in an environment that does not reward your efforts. And I know there's like loyalty and all, all of these different things. But I can tell you just how refreshing it is to walk into a room with someone who thinks the same way you do. And I can promise that you will be rewarded so much more in a firm that aligns with your way of thinking than the firm where it is an uphill battle every single day you clock in to work. If that's you, get in touch. So the company I acquired is Cloud Account Staffing. I'm now a co-owner there and we've got a whole team of people that are now making these placements for people exactly in that situation. And for firms who are hungry for that type of person, a more forward-thinking accountant that is kind of thinking in the same way about how to run an accounting firm. I mean, imagine going to work for a firm where they go to my live events and listen to the podcast and all that stuff. If you're into this stuff too, imagine having that in common with the firm that you go to work for. I mean, it's, it's life-changing. That's why we're doing this stuff now. If that's you, get in touch or head over to Cloud Accountant Staffing. Really cool team of people that run that company. It's the only reason why we would have even considered doing something like this. And is like a pretty core part of like the mission of like what, what are the, one of the most life-changing things we can do with this business that isn't just make another silly YouTube video. It is pulling people out of really bad situations and putting them into a situation that is life-changing, a vastly better job, and honestly, even keeping people in the profession who otherwise don't know that there's an alternative or don't know how to find their way into a better firm. If that's you, we can absolutely help with that. Gang, You know what? I think you're making the world a better place just by showing up here, just by liking and subscribing, just by leaving this podcast just a 5-star review. You're making the world a better place just by doing that. All right? That's what I want you to think so that you come back and tune into the next episode as well. I can't wait to see you there. Until next time.