When Financial Reports Used to Be Cool & AI's Microwave Moment

The staff are gone. It's managers, directors, and partners. And the firm of the future will basically be a team of those people overseeing a bunch of AI agents and maybe some administrative. Coming to you weekly from the on-pay recording studio, hey everyone and welcome back to this show. This is your weekly round up of news in the accounting profession. I'm Blake Oliver. I'm David Leary. Blake, I'm gonna I need to talk about some experiences that with AI this week. But first, let's think our sponsors. Sponsors this week we have Cloud Accountant Staffing, Numeral, On-Pay, and Veltrix AI. So two new sponsors should be sure to stay tuned to here are their ads today. Are you tired of the endless shirts for qualified accounting talent? You're not alone. Growing accounting firms are struggling to find available and affordable team members when they need them most. Cloud Accountant Staffing has the solution with their revolutionary candidate portal. Unlike traditional staffing agencies that waste your time with sales calls, paperwork, and deposits, the Cloud Accountant Staffing candidate portal gives you instant access to highly vetted qualified accounting professionals. No waiting, no hassle, just top talent right now. What makes this different? Speed and simplicity, while other firms make you wait weeks or months with Cloud Accountant Staffing, you could interview someone as soon as tomorrow. Their boutique support ensures you're getting quality talent that's both available and affordable. That's exactly what growing firms need. The candidate portal puts you in control. Browse live candidates, make selections on your timeline, and build your offshore team without the traditional headaches. To find review and book interviews with potential team members, all in less than 10 minutes, head over to accountingpodcast.promos/casts. That's accountingpodcast.promos/cas. Yeah, but I had two experiences with AI this week. One, I'm kind of a ho-home opinion about, but I do think there's a warning for accountants here. Then a second one, I had an experience that we're going to with you and AI. First off, I'm turning 52 next week, right? You have to do your yearly physical. Are you at an age where you're doing yearly physicals and blood draws every year yet? I should be. I don't like going to the doctor. I got my blood drawn, and then they send to the results, and all these doctors have these portals and they send you an email. I got an email, and I think this doctor's note might be AI, because it was just too generic. It said, "Hello, your cholesterol numbers are still pretty high. Dash not urgent." It had the dash, right, the tail-tail sign. But you can discuss with Emily at upcoming appointment. Diet, low fat high fiber, at least 30 grams of fiber per day. Exercise, 30 to 60 minutes a day. Healthy weight. He knows, I literally go to orange 30 classes with my primary care physician. He knows I get exercise. He knows probably how much I weigh. I'm pretty sure AI sent this. My thing is, I fear this how accounts are going to use AI. You're busy. I'll just have a reply to the client. It's just going to give a shitty pointless response to the client. The client's going to get a bad experience. It won't feel human. It'll feel like why did why did I hire this person just to send me a template and email? Exactly. I don't do this to your clients. Treat your clients with the time and energy they deserve. If you want them to pay, you can do it with your free clients. But the doctor, you pay a lot of money for. The second thing this week, AI that's really got under my skin blink, it was you. Your AI went into my database schema for all the ads you do for the show and changed the database schema in our database. Arguably it was minor because I was able to fix it quick. But if your AI though didn't brag about doing it in an email, it sent me. I would have never noticed, probably for another two weeks, when math didn't work or something. I made a mistake in mischaraculate or overpaid somebody and then probably blamed our human salesperson for changing the database. It's crazy, right? David, your big mistake was giving me administrator access. Yes, I've downloaded your access. No, I'm just a lowly editor and my AI plugged in acting on my behalf couldn't actually add a new field value, which apparently is what it did because we were trying to mark some adslots as pending for 2027. It just created a whole new status of pending. I mean, it's a good point. I apologize, David. I was, you know, touching playing in your sandbox there and went too far. But I mean, it's a good lesson in internal controls, right? Is if you're going to give access to an AI agent to a system that you use, you need to set up a login or permission so that it can't do stuff like that. Like the same way of human. You wouldn't want a human mucking around doing it. Yeah, because the humans I've walked down, but I thought but because I think the thing is, Blake, like you would not have done this if it was you going the database, clicking around, you pilot a bit. I probably should ask David before I just edit the field settings. So what I did for my own accounting is in zero. I've got an AI agent doing the close in zero, the month in close. I created a login under our practice manager account just for the AI agent. It's called agent and it's agent at and that's email address. And then that's the one it logs in with. So we can actually see what it's doing in the audit trail, whether it's going through the connector or through Chrome. And it doesn't have administrator rights. So that's on getting around that. Yeah, to lock things down. That's actually a good tip. I mean, create setup. If you're going to set up agents to be on your team as little many humans, or if you want to call them, give them an inbox, give them a login. That's just them because that's the best way to audit it. You can see how they logged in. They made this and then you control their data access. Well, David, I've got a video here. I want to play for you about financial reports and about when financial reports, the companies put out like their annual financial reports used to be cool. And I didn't know this because I'm from a different era, I suppose. And now they're all just super boring, right? You download them from the internet. And it's just a bunch of texts. And maybe if you're lucky, there's some images. But this is kind of a cool overview of what it used to look like. Corporations used to treat annual reports like modern art, but today it's just another spreadsheet. Pre-internet, these reports were the only touchpoint businesses had with investors. So it had to communicate success, innovations, stability. So they went all out. Beautiful graphic design, expensive paper, bespoke techniques. It was the corporate flex to make a coffee table book level report. Then two things happened. Burst the internet made it so the physical report was no longer necessary. Then the 2008 financial crisis spawned strict guidelines on the way this data is presented. So in the eyes of these companies, it became too much of a risk to get cute with the design. But the story doesn't end there because we're seeing a bit of a renaissance in this space because this random data storage company realized people pay more attention to data when the design around it looks good. So now their entire content strategy is just making mock-up designs for earning reports. And they do respectable numbers on X. The story is exposing a fundamental truth about human nature that too many business people ignore. And that's that if you want people to pay attention to something, the first thing you have to do is make it fun to look at. I don't think it has to be fun. It's the experience, right? The investors are the client of that report and they gave them something they would have a good experience with. Right, it wasn't. Yeah, and it's good reminder, you know, like for accounting firms, like the reports that you deliver to clients, like make those beautiful, make them want to look at them so that they're more likely to open them and not just disregard them sitting in their email. That was Lucas Mullin on Instagram. All right, David. Now onto the news. I've got a story about taxing AI. As we've seen AI agents and LLMs improve over the last few years, there's been more and more talk about taxing AI because these AI's are going to put humans out of work. And now Bill Gates is out saying that he wants to tax AI tokens and robots. His reasoning is that if you hire a person, you pay payroll taxes. If you buy a robot, you get to write it off as a business expense. So there's like an economic disincentive to hire people and an incentive to hire bots and depreciate them. Just tax policy dictates social policy. So on a recent blog post on Gates notes, he floated this idea that he calls human reserved. Certain jobs would be set aside for people only either because automating them would displace too many workers who can't easily retrain or because a human should be doing it. And his example is a robot telling you that you have an incurable disease. It's technically possible, but it shouldn't happen. And on speed, he says the usual comparisons about AI and other technologies don't work. The PC, it took about 20 years to really change how we work. Software had to get built. Prices had to fall. People had to learn to use it. But AI already runs on the devices that we use and it speaks in plain English and we don't have to adapt to it. It adapts to us. So he thinks that areas like law, customer service, medicine, software, manufacturing are going to get hit over a decade, not generations. So like the change that we saw from the computer revolution in completely rewiring our economy from manufacturing to information, that's going to happen not over generations, but in the next decade. And one of the examples that he gives is an accounting worker replaced by a bot in this, in this article. And you know what's interesting about this is accounting is always on the far end of the chart when it comes to jobs that could be displaced by AI. But we have not yet seen meaningful job losses. I've waiting for one. I don't know one yet. One person in accounting that lost their job because of AI. So I feel like there's some sort of disconnect here between what these AI companies are saying can be automated, eliminated, and what's actually happening. If accounting is so easy to automate with AI, then why aren't we seeing accountants lose their jobs? Maybe it just hasn't happened yet. Maybe it's coming. But there's a part of me that thinks that it's not. And it's because of some, something that these technology people just don't understand about. Well, of course, it's the same mindset when they build the GL. Like, oh, I have a startup accounting. So easy. I'm going to build the GL. Well, it's not good for plumbers. No real businesses can use it just other startups. They have this, they have a broken view of the world. I saw another story that maybe gives us a clue as to what's going on. This was in business today. And it's a headline is US jobs paradox AI exposed roles see 46% pay surge even as AI linked layoffs rise. So companies are using AI to automate tasks and reduce headcount. But the workers in the most AI exposed occupations like accounting are actually seeing the fastest pay growth. That doesn't make sense, right? You would think that in a accounting, we would be seeing pay go down if AI is competing with us now. Yeah, if it's actually taking the jobs away. Yeah. So the takeaway here is that it's not necessarily a bad time to become an accountant because what we're going to have is leaner teams, but higher compensation. And since we have already been in a talent shortage situation for several decades now, this has gotten worse and worse, we are actually perfectly positioned to take advantage of it. We don't have enough accountants. We're going to have to use AI to do more with fewer people anyway. And there are going to be compensation bumps higher than the rest of the economy for accounting. That's my theory based on what I'm seeing here. But yeah, it's funny how like all these people always use accounting as the example in their stories. Yeah, I guess because people just don't, people don't value it. It looks easy. It seems easy, but it's not. One more story here related to all of this. The head of the PCOB was at an Axios live event recently. And his name is Jim Logo Thetis. He's the chair of the PCOB now. And he said that when asked about what's going to happen with audit and with AI, he said that audit is part science and part art. And that the science and the the transaction work that can be handled by the machines, it's already happening. But the art part of accounting is a different story. Professional skepticism, judgment, knowing what you're actually looking at in a particular industry. And that part according to Logo Thetis needs people still. And so AI is good at those things. I don't know. Well, some of them, some of them, right? It's like, it really depends. It depends on whether or not you can create a strict criteria for success. So if you can define and write down very, very detailed criteria for success, the objectives, how to do it, AI at this point can now follow an instruction set. It can follow a process and it can reliably get to the objective and correct itself along the way if there's enough context, enough information. And it does that by using lots of sub agents and creating checklists and working through the checklists and converting probabilistic outcomes into yes or no outcomes. So if one sub agent fails or here's an example, this is how you can really improve the quality of an AI prompt that you're writing is like in cloud co-work now, you can explicitly tell it to spawn sub agents. So you can say to do this task, I want you to spawn two sub agents. And I want them to, they have to agree for this to be successful. They have to come to the same result just like two humans checking their work. So a secondary table in Excel, you're doing a check sum against the other values. Exactly. So let's say you have a 90% success rate for a particular task with AI. If you have two of them do it, that reduces the probability of failure dramatically because they'd have to both hit that 10%. So I don't know if I'm doing the statistics right, but it would be like 10% times 10% is like now there's a 1% failure chance. And if you want to reduce it even further, you have three AI agents do the same thing independently. And that's how you get the accuracy up. You can increase your token spend. So you're saying if you just a toss more token spend and more agents at it, now you can get to six sigma. Six sigma? What do you mean? Well, six sigma, you might be in that weird generation that like it bypassed. If you're older, you know, a lot more six sigma. It was the GE way. And it came from Motorola originally. And to be honest, it almost killed into it at one time. It's all about perfect data, perfect measurement, perfect numbers, but the six sigma is like you're uptime is 99.999, but that means you're down for one day a year, right? But essentially, that's that's the which basically kind of what you're saying, right? You could get from 90% quality to 99.999 if you just put enough money and tokens and agents on it. So what the PCOB chair said fits in with this in the sense that the science about it, the stuff that you can actually document down to the smallest task, that's all going to get automated with AI. It's the judgment part above that that won't. Well, what's left when you do that? The staff are gone. The staff are gone. But how are you going to create an informed opinion from not in the weeds if you're not eating the dog food? Because those people already aren't in the weeds. They've got humans doing it for them. And just they're getting reports. They're getting spreadsheets. They're getting, you know, they're not actually get the ability to have this judgment by doing the work. And that's the challenge, right? Is how do you make more reviewers if you don't have preparers? And nobody's figured that out yet. This professional judgment comes from eating the dog food and doing the work. It's not, I mean, obviously, who set up their training simulator last week? We talked about it. Was it Y? KPMG, I think. You have to use years of this simulator before you get to that point of having professional judgment. This is why it's called professional, right? You're practicing your profession. All right, David, let's go to our next sponsor. And I'll let you take this one. This is a new one, right? Numeral? Numeral, yes. Let me ask you something. When you recommend a sales tax solution to a client, are you confident in the support they'll get? What about the pricing? What happens if something goes wrong? As accountants, you've, you've seen it all, clunky support, surprise invoices, and nobody picking up the phone when there's a problem. Numeral is built differently. It's a modern sales tax platform built for accountants and it supports registrations, filings, nexus tracking, exemption certificates, even vat and GST in over 70 countries. All with great support, simple transparent pricing and no surprise invoices. Then there's the guarantee. If your client's sales tax isn't filed on time, Numeral pays the penalties in the interest. And for accounting firms looking for a true partner, Numeral works alongside you to support your clients as they grow. Whether implementing it yourself or making a referral, Numeral has you covered. To check them out, head over to accountingpodcast.promos/numeral. That is a accountingpodcast.promos/NUMERAL. Since we're talking about the big four, let's talk about one of them in particular, EY. EY has reported improved audit quality, according to accounting today. It's the biggest turnaround in PCAOB inspection history for the firm. Back in 2024, 18 out of 64 audits reviewed had significant deficiencies. That was a 28% deficiency rate. So over a quarter of the audits that the auditor of the auditors inspected did not pass inspection. In 2025, it was just 3 out of 64. That's under 5%. So they went from 28% down to 5% in just a year. [BLANK_AUDIO] And they are crediting the $1 billion investment into tech and talent, including agentic AI tools that every US audit and tech risk professional can now access inside of their audit platform. I feel like we talked about another firm that hadn't improved it. It may have been BDO or somebody they said, "Oh, it's because of all the money we spent on tech." But they had a big improvement, too. So is this true or is the PCAOB just not measuring in the same way as they used to measure? I don't know. Well, I guess it depends on what the PCAOB is actually finding fault with in these audits. And my guess is that the PCAOB is traditionally very nitpicky about actually following all of the processes that you're supposed to follow in the audit, not skipping procedures, having all the documentation and all that stuff. Well, guess what is really good at making sure that you don't skip procedures and that you have backup for everything, it's this agentic AI stuff, right? All that just checking the box that gets missed when people are rushed. And speaking of people getting rushed, there's another data point here from their press release that says something important about how to actually improve audit quality. 69% of the public company audit hours at EY, so over two thirds of those hours are now done before your end. They are not having to wait until after to cram them all in. And so this makes sense. If you have less busy season crunch, you're going to have less rush judgment, fewer managers and directors partners rushing through and not able to review things properly and push the work back down and just, you know, pushing it through to meet the deadline, you're going to get a better result. So it's not really AI, the use of AI is making it get done with higher quality. It's AI is changing, it's spreading out the workflow or the workload across the year. So people, the humans can function properly and not make mistakes on the audits. And I think it's probably both and we don't know exactly how much. But given that back in, well, only a year ago, there was not nearly as much agentic AI stuff going on. You say, "I agents weren't as good." I think it's probably the pulling the work forward that's doing it. And theory, the PCOB should use AI and audit more of the returns and not just take random samplings of what, 30% just, actually, that's a great point. PCOB, right now, they just select a certain portion, a very small portion of the audits to inspect, well, they could just do all of them. They could say every audit firm has to upload all your work papers and everything into some sort of portal and then run an AI that examines all of them and looks for issues. I wonder if anyone over there is thinking about that. I bet the big firms would love that. I'm sure. I want to jump into AI had it's microwave moment this week, Blake. What is it? What is a microwave moment? So, think about, at one time, you had to know how to cook, right? You had to have some skills to cook something and then the microwave came out and as long as you could push a button, you could cook food, heat up food and eat it. Two-year-olds can cook with a microphone, 99-year-olds can cook with a microwave. Like this is the moment it had this week because Meta dropped its new AI called Muse and right away it was the number one downloaded app on the Apple App Store. I was like, "I should try this out. Let's check it out." So, in part of their press release, they say there's no learning curve. Anybody can use it out of the box, no technical experience required. And I used it because I'm going on a trip to do some research at parties that are at bars before the bills game in LA. And I sent you the link. I was like, "Oh, my God. Look what this did." And I could say, "Oh, just tell Claude to make an artifact and he'll do it." But the point was, I didn't have to create projects. I didn't have to configure background agents working. I didn't have to ask for an artifact. I didn't even know I wanted one. It made this beautiful thing, this output that I didn't even know I wanted. All our moms Blake are going to use Facebook Muse. This is the AI for our moms and those Boomer partners, right? Yeah. Because, sure, they're probably dabbling with AI right now asking Google search question, you know, asking a question here and there. But they don't really do stuff with AI because they don't know how to configure it and set it up and do all that, all the stuff, or even know what they want it to do. And so this really, truly is the microwave moment. Like, anybody's going to be able to use AI because of this. And what hopefully will happen is the other models will copy this. Right. Right? Like, honestly, it takes, it's on your play, like, "Okay, I've got to go set up a project. I got to link it in." And that's all, like, brain waves, you have to spend to make sure it's all configured properly. Like, out of the box. Yeah, most people. It's just amazing. It's amazing. Yeah, most people aren't going to do that. Yeah. Okay. I mean, that's interesting. I wonder how much meta is going to spend on Muse because creating what you sent me, that, like, Bill's schedule, that beautiful artifact that it built or webpage essentially, like, that's going to be a lot of tokens, right? And so if people are, like, messing around with this, the free, at Facebook scale. Yeah. And so this is the issue, I think, for Facebook is they're investing all this money in AI infrastructure. But their model, their business model, is free. And tokens are expensive. So how do they make money on this? That's interesting, too, because it'll go shopping for you. They're using, like, a stripe payment portal. But they insist it's siloed out in its own virtual machine. Just for you, it's Facebook's ad network can't see what you're doing and insert ads. But I don't know. I don't trust Zuckerberg. And I'm kind of at that level with some of these AI companies, you don't trust them. But they claim that there's no ad integration. But that's a good question. How are they going to make money for this? So they are. It's free for basic use. But they are going to have paid subscriptions for heavier users. Now, one of the things they said, oh, to book your hair cut. So I tried to have it book a haircut for me, because I need to get a haircut. It couldn't because my haircut, please, only takes phone calls. Now, for if you're a firm and you don't have a place to book an appointment on a website, and I tell them, use, hey, book an appointment with my accountant, it's not going to be able to do it. You are going to lose business. So you need to make sure you have your website and your inbound to your firm set up in a way that AI can go book things. And oh, David, you just gave me an idea. What if an accounting firm had their own MCP server? What if you could plug in to the accounting firm directly through whatever AI bought you're using through mues or cloud or tattooed PT and connect directly into the firm to submit documents? No more portals, no more emails. Yes. Actually, even beyond that, right, hey, mues, go get all my tax stuff off my Google Drive, go find it and just put it, just give it over to the accountant. Right. You're going to have some interface for machines, if you want to call it machines or bots through agents to your firm. You're going to have to. The same thing with paying, because it can pay for things, it can go shopping, you can pay for things. So you can have a form where it books the appointment for me and pays you all through my agent. And I mean, they claim all this privacy stuff, but I don't fully trust them. I didn't connect it to my email or calendar. Oh, yeah. I don't trust Facebook at all at this point. I just paste it in like a screenshot of my calendar appointment. Like, hey, I'm on this airplane flight. Here's the game time. Go find me all the bars that are events happening at. I like this, though, David. This is how we get the boomers onto AI. For good or for bad. It's going to be Facebook. Yeah. All right, this is how you get people in your firm that don't know how to maybe just they use the Google prompt and the doai in there or what's built in the Apple phone. And that's about it. They're not doing the next level because it's you got to understand a little bit on the real ship of a project and how it ties. And one artifact is you don't have to know any of this anymore. Let's think our next sponsor and that is on pay. Are you tired of payroll headaches getting in the way of the client experience you want to deliver manual workflows creating bottlenecks compliance nightmares and endless support calls that go nowhere. There's a better way for your team and your clients on pay is the payroll partner that accountants and bookkeepers actually love why because it's easy to use packed with value and backed by support that actually supports you. Their team gets rave reviews for being fast expert and actually reachable when you need them on pay handles the heavy lifting. You get a dedicated onboarding on boarding coordinator who sets up worker profiles and transfers your to date data from previous providers all at no extra cost. There's seamless quick books and zero integrations eliminate manual journal entries and they support any type of business you serve farms restaurants nonprofits you name it on pay can handle unique requirements without adding complexity and on pay keeps pricing simple to everything your clients expect from multi state filing to off cycle pay runs is included no hidden fees and no surprises to book a demo head over to accounting podcast dot promo slash on pay that's accounting podcast dot promo forward slash oh N P. A. Y. So I have two two articles that are tied together it's something expensive I did something zero didn't ties back to a conversation. that you had with me and your opinion on things. So, Expensify announced that they are one of the first Expense Management tools available in Clod. And so, you can just inside of Clod, now, ask questions like, which expenses are missing when I receipts and not actually leave Clod. But, I'm starting to feel like this is kind of dumb because, I saw this as like, what about the new Expensify? Can I just do that there? So, I opened up Expensify, which is essentially arguably a chat interface. And I said, exactly that question. Which expenses are missing receipts? And it said, "Hi, David, you can only have three expenses that are missing receipts. If you'd like it, can help you show which ones those are, or help you get them cleaned up." So, basically, Expensify is training users to not use Expensify. Now, think about this. There's risk to this. And, to it's doing the same thing, and I'm going to talk about zero here as well. You're training your users to not open your app. And that is probably not safe long-term. Why? Do you remember, think about this. We went, just like, I feel like it was yesterday. We went from apps adding chat interfaces inside the app to now telling you, chat outside of our app. Remember, Jack's from zero. Is that even a thing now? Because zero is now launched, an integration with chat cheapety and cloud letting their 5 million small business users open these, get data in cloud and chat cheapety without app switching. Right. And this is the thing, and I think you said this too. The value prop they're solving is, I don't have to open a second browser tab. What is the point of this? You're giving up control of your data, your users. It's really risky for zero into it and to intensify in these other companies to let cloud own the customer relationship. Oh, be the interface. But here's the problem, David. If they don't, somebody else will do it. And I am going to switch, because I don't want to have to go into a separate app and do the thing. When I'm working with an AI agent that's closing my books and helping me process my spend, I want to do that there. I don't want to do it separately manually. And so like, that's why RAMP has an MCP. And I can provision virtual credit cards via a cloud chat. I mean-- So this is going back to my theory, which is the lesson we learned about Microsoft for Microsoft Windows back in the day. Microsoft is in bed with Apple, IBM with OS2, and they're building Windows itself. Whoever was going to win the OS game for your computer, Microsoft was going to be a winner. So that's kind of your opinion, like they have no choice, because if not, cloud's just going to build expense tracking right in the damn app. Well, somebody will build it and integrate it. I think that the way to think about this is that cloud and chat GPT and Gemini are the new OS's. That's the user interface. And all these apps are plugging in just like apps you used to install on your desktop. And so whoever grabs the audience, whoever grabs the users, is going to be the OS layer. And with cloud, it seems like it's professionals, and especially accounting and finance people. Cloud has really taken that market, the developers too, and owns it. It stole that from chat GPT almost overnight when they release co-work and cloud code. And chat GPT kind of tried to catch up in this kind of field. And chat GPT has actually embraced the consumer market. Most of chat GPT's revenue is those individuals, not organizations at this point. And so there's going to be, again, we're going to have like two OS's, right? We're going to have the, it's like Apple versus Windows. Or yeah, and maybe a few others too, right? If they can catch up. But it's interesting, because I'm having to revise my whole theory about how these companies are going to do, because I thought, and this is, I still stand by this, that if you're just building large language models and selling access to them to tokens, essentially, that's a commodity eventually. Because there's going to be many, many, many good models. There's no, there's no IP on any of that, right? Other than a specific model, but anyone can make a model, right? It's not, there's no patent on this technology. And so that's going to go towards the cost of electricity. So where do they capture value? Like where do they create value? It's in that management of all of the stuff going on, management of the agents. So cloud co-work is so powerful because it allows you to organize all the work that your agents are doing into projects and scheduled tasks and skills and maximize the effectiveness versus just single prompts. And you can do really complex work now because of it. So that's the value add. It's the orchestration layer. Yeah, I'm going to talk about orchestration layer next. I'm going to start, but I want to close this thought on this. Ultimately, our cloud code was actually another company called Codex. And they've learned what they're doing in this basically stole it. And that's the danger of this. Expensify might not exist in a year because cloud will just steal it and just do it. We'll just see cloud expenses. And you'll just do what Expensify does. So why would you prematurely send your people into those apps to do these things and let these companies learn and steal from you? Because that's exactly the Cloud code. They're doing it to Figma. All the developer tools they're doing this too. They're sucking in and just stealing the customers and stealing the business. And Microsoft did that. Microsoft was either an Apple did that. They acquire your app. Yeah, history of platforms. I agree. It's a history of platforms. And all you know, we know the history of platforms, why would Expensify and zero do this and into it? It doesn't make sense. They're going to get screwed by these platforms. They can't trust it. Well, they can't be trusted, but also you know, you're not the low-hanging fruit for them. Creating payments rails and routing like payments through the banking system is not something that the developers of cloud want to do and build because that takes a lot of time and there's a lot of regulation. But if all you're doing is like creating, I don't know, designs in some software, right? That's something they could clone. Writing code, yeah. And something with a good GL, right? It's like creating a multi-tenant general ledger system that can serve millions and millions of businesses, something that took into it like decades and decades to build. Not something that AI companies are going to come after any time soon. And go ahead if you were the orchestration layer, right? So Avalor just announced a new product called Avalor Aviator and that is the orchestration layer for the AI agents. So if you think about they have the orchestration agent, I think it's called Avi, and then you have a transactions agent, a returns agent, a research agent, a certificates agent. So these agents are doing all this work rolling up. And this has kind of, has been my thoughts, I want Dem AI, right? I might want, I want AI that just does one thing really well. All you do is enter bills. Another AI that does nothing but create invoices, right? So the smaller is, and then I want a parent AI agent or the orchestrator directing these things what to do. But I don't want my agents having knowledge of the entire world. Now I was thinking about this and in theory, shouldn't it into it and zero be building something like this? You have hundreds of apps that are trying to write to the GL, right? And nothing's controlling that. I mean, there's an API call and an API key. But the data itself and what gets written and how fast and is it correct or not, there's no. Yes, I think there's an opportunity for, into it and zero, to be the quote unquote, to build an agent that's like the human in the loop for the GL. And this is something I think way back in the day was that into it, we had beer conversations about this. If you think about the GL, right now in the world, the API world of cloud accounting, if whatever third party app developer had a transaction, they never had all the details of a transaction, right? But they have some parts of a transaction, and they have their parts over here in the strike, the payment part and this part's over here. And from an API perspective, these are beer conversations a decade ago. Imagine if the developers, instead of sending a perfect API call and they could just send half a transaction, I know this about a transaction, and then there's an orchestration level at and to it level that wraps out all up and then writes a real transaction to the GL, right? There's part of an opportunity for, because you think you said this orchestration level for them to control the data getting into the GLs. And that's what I would like them to see them do. Here's the thing is that orchestration layer would be limited to accounting. And I'm gonna use an orchestration layer that is unlimited to anything, this general purpose. And that's why like cloud co-work is so powerful, because I can use it for almost anything. And my entire business can be in it, run by it, not just accounting, also sales, also marketing, also operations, also production, everything. And that brain, that mind and all those agents that are inside of it, which I've been building the last few weeks is wild. You don't want that broken up either. So, you got one universal chat. That's your one interface to the whole world of data and all that. whole world of everything. You don't want to open browser tabs, you want to have a different chat program. You want one interface for us to your life, is that kind of thing? Yeah, and if I have to open browser tabs, it does it for me and tells me where to go. It drives for me. So I can look at, know what I'm looking at and get there automatically. It's really incredible. Hey, David, we need to thank our next sponsor. I'm going to let you do that. And then after that, we're going to get to some of the live stream comments. You're on. You got on pay? I did on pay. You are on Veltrix. Veltrix. How much of your week is spent pulling reports, cleaning files, explaining numbers? Clients still don't really understand. That's exactly where Veltrix AI helps. Veltrix is finance as a finance AI co-pilot and collaborative workspace for accountants, advisors, and clients. Each client gets one private workspace where QuickBooks zero and a bunch of other sources, messy PDFs and spreadsheets can all safely sit together. You ask a question. Veltrix gives a clear answer and the actual reports and the sources behind it. So your team and the client can work from the same context without exporting data just to double check it. And it's read only so it never changes the books. Veltrix does a first pass, flagging duplicate charges on the usual entries and slow playing customers before you go looking. Firms also get onboarding AI education and direct access to the team behind it. If you want to safer way to bring AI into client advisory work and to get one month free by using code free month 26, head over to countingpodcast.promos/veltrix. That is the countingpodcast.promos/VELTRIX. Welcome to our live stream viewers. Thanks for catching us live. Mail 22 says people are not getting fired yet, but definitely not hiring new ones. So the hiring growth has slowed down. Yes, that's true. We'll see if the firing happens. El Capitan on X says, how does the future of Deloitte look in comparison to the other big four? That's a great question El Capitan and I have no clue, but I asked Claude while David was reading his ad. Deloitte is, you know, enormous. It leads PWC by 17.6 billion, which is bigger than most firms entire consulting practice and it's also got a massive consulting practice and that's been driving most of its growth over the many years and decades. But that's also now its weak spot because AI is driving down demand for consulting. AI is really good at making all those power points, all those reports you got to use to justify whatever it is you want to do to the committee. Technology and transformation deletes biggest consulting business, grew only 2.5%. Last year, the previous year, it grew 4.7%. But the slower businesses grew faster, tax and legal, audit and assurance, both above 5%. And it's interesting, this consulting thing. I really do think that as AI's get better, more and more executives are just going to start using it for all that stuff that consultants were doing because there's got to be truth behind the joke that consultants just reuse templates and come up with meaningless notes for the slides and say a bunch of stuff but they don't really mean anything, right? And isn't the joke that a lot of times when you hire a consultant, it's like CYA, right? You're covering your ass, right? You want to do something and you need to hire Deloitte to give you the cover to do it. And if it doesn't work out, you just blame Deloitte. Well, I mean, maybe that won't go away, but it's still going to get done a lot cheaper, right? And it's going to put pressure. And AI is putting all this pressure on consultants to do the work faster. And I actually have a story in there in my queue about that. It is from Business Insider. Headline is clients now expect some projects to happen twice as fast. This is actually a PWC partner who talked to Business Insider. Janelle Johnson leads PWC's 2200 person DC office in consulting and says that a project that used to take eight to ten weeks is now expected in about four. She says that the tools should accelerate the work. That's the expectation from the clients. McKenzie says AI turned its week one answer into an hour one answer. A Deloitte engineer said some coding tasks dropped from days to five minutes. An EY consultant said due diligence that took two weeks of spreadsheet work now takes a day. So all of this stuff that used to take a lot of time is now taking less time. And here's the question that nobody answered in the article. If the time shrinks, what happens to the bill? And I don't think that bill can stay where it's at when the clients are seeing it take half the time or less. They're not going to put up with that. So the fees are going to have to shrink on these consulting engagements. Unless you're somehow delivering more value, it's not going to work. The fees are going to come down. And more and more smaller firms are going to be able to do the same quality work that Deloitte and PWC did today or in the past. So I'm glad you asked that question, L. Capitan, because I think the future is often contradictory. It doesn't match what you hear. What you hear in the press and the mainstream media is that accountants are screwed. And in reality, it's the accountants who are going to be making the most money if they adapt to using all these AI agents and move up the chain. And it's the consultants who are being highly paid to do theoretically, you know, sophisticated work that requires a lot of judgment, but really it's just PowerPoint, you know, mucking around. Those people are going to lose their jobs. And we all know those people. I have plenty of friends from college who went into consulting, and they're not geniuses. And this sounds like a career at big four kind of question from L. Capitan, right? That is looking for. So I have some career advice. If you are thinking about going to big four, or if you're already in big four, if you're at KPMG, that's where you want to be. And here's the reason why. So obviously, we know KPMG and all the big four, they've all announced they're going to spend billions of dollars. Well, KPMG just announced that they're launching. It's a horrible name. It's called the client technology innovation group. And essentially what this is, it's going to be like a lab, right? It's an incubation lab where they're going to have rotate different people in from the company, rising stars are going to roll in and they're going to build AI come build companies, build apps, build like a startup accelerator, right? They're going to build new products and then roll up from there. And they'll work closely with Anthropic, Google Cloud, Microsoft, but they don't really want to build like a lot of AI type products. They want to build out. And arguably they have no choice but to do this. They blew or blowing over a billion dollars right on AI. They need to figure out how to create products that they can grow and make money off to earn back all that spend. And it sounds great on paper, right? Rising stars will rotate through the CT and I as founders and leaders of edge companies, building AI literacy and entrepreneurial skills before returning to core leadership roles. But I don't think it's going to work out this way. And this is why pay attention. This is a good career move. What happens with these labs at these corporations? There's corporate bottlenecks. So the people that are in there, the founders, these people like that, they usually leave and build very successful companies. And I've seen this happen when I was at Intuit back in the day, arguably me, I was really heavily involved in Intuit labs. And it let me like, you know, sharpen my skills with startups and creating companies without the risk, right? I was getting a nice paycheck from Intuit. It was safe, but I was able to launch new products and build stuff when I was at Intuit. And that was in the other people that were there. Webflow was came out of that was a former employee that was at Intuit labs. They launched Webflow at one time, hit a four billion dollar evaluation. There was an app called TEMPO, which is a contractor and payment app that was built by people that were formerly in the Intuit lab dot dot bought by Stripe. But if you go further back in history, yeah, who used to do this? It was called the Brick House. And it was an internal startup incubator. And then people just got frustrated and left. Well, you ever hear of Y Combinator? That is former people that were in the Yahoo Brick House. What's that? Former people that were in the Yahoo Brick House. So if you're at KPMG, you should try to get involved in this brush up on your entrepreneurial skills while you're on the paycheck of KPMG, get good at launching and building products, make mistakes, then go start a company. So what's going to happen is KPMG is spending billions of dollars to do this, right? And we're going to see a few billion dollar startups that are tied to KPMG from a resume standpoint, but not actually tied to KPMG because people are going to go to good idea and they're just going to quit. They're going to quit. So it's not good for KPMG, but it's really good if you're employee KMG to get KPMG to get involved in this incubator. Do it. It's a really good idea. We don't have any more sponsors, right? So we can go on to our next story here. Yeah, we can talk about hacking in AI. If you want that, FBI attack. I want to talk about how America has become an entrepreneur's paradise. And this is from the view over the Atlanta Atlantic from the UK. This was a story in the economist. Why is America now an entrepreneur's paradise? I mean, I guess we kind of always have been, but it's getting even better. This is the good news. I mean, not feel that way with everything going on in the world, but it is. Here's the numbers to back it up. Americans filed over half a million new business applications in June, 531,000. That's double the monthly average from 2019. Okay, half a million new business applications in June alone, double the monthly average from 2019. Since the start at 2021, we have averaged 466,000 applications a month. And the share of founders who used AI to help launch their business doubled to 60% between 2023 and 2025. That's according to gusto. And it makes sense. If you're using Claude every day or chat GPT every day, it can help you do all of this crap that used to be so hard and time consuming. You want to build a website, get a license, figure out the rules, do back office work. You just have to do that yourself and you have to do that notes to the corporation commission and all that stuff with the corporation commission. I mean, you can basically have AI doing all that stuff that's not core to the business that you're doing. Now, here's what's interesting is that these new businesses are less likely to hire people. Only 30% of the new applications last year were high propensity, meaning likely to hire at least one person beside the founder. And that's down from 38% in 2019. So an 8% drop. So less than a third of these businesses are likely to hire more people. That means that, you know, two thirds of them, right, 70% of them are just going to be solo. But you can do so much more now as a solo printer. So I think we're going to have a renaissance of self employment, single person companies. Because you're going to be able to do a lot of stuff that used to take a team without having to hire a team. And that's why I think like the CPA firm of one, bookkeeping firm of one, tax shop of one, audit firm of one, consultants of one are going to be huge and they're going to make so much money. You know, instead of making like low six figures, you're going to be making half a million dollars a year, a million dollars a year. I agree with you and I love this concept of the company of one, right? I really love this concept. Now, the thing is these business filings don't necessarily, I don't think actually believe they're tied to real businesses being started. We saw this with COVID, when the, when the stimulus money came, all these new businesses just started getting created because people were trying to get COVID money, right? You need, and now think about all the Instagram influencers. Oh, you need to start a business and LLC and then put your car pink slips in there. So you don't have to pay for taxes. How many are actually real businesses being created? How many of it's just business structures to because it's the thing to do? Like are they, are they actual businesses being started? So that's where this, this word entrepreneur, just because somebody applies for a business license doesn't meet entrepreneurship's happening. Maybe that's my point of view. And I would argue there's probably a lot of just paperwork happening, not actual entrepreneurship. All right, David, I'll let you take the last story. I think we both have two music stories we should touch on. Okay. What's yours? I know who wants to go first. I can do it. I can go first. Okay. So this is a op-ed in accounting today by our friend Giles Pearson. What accounting firms can learn from the music industry? And he says that what Spotify did to music, AI is doing to accounting, which as a musician, I don't know, that worries me because the music streaming service is basically demolished the music industry. When I was just getting into it, years ago, getting out of college kind of screwed it up big time. It's changed now completely, right? It used to be you made your money from record sales. And now it's not none of that. It's all merchandise and touring and all that stuff completely shifted, but that shift was really painful. Anyway, I digress. So here's what Giles' argument is. He says that AI can accelerate tax returns, compliance reports and workflows while cutting costs, but your competitors are going to get the same benefits. Clients also have AI-supported advice tools and more ways to obtain work that was once available primarily from accounting firms. So his warning is that adopting AI itself is not a strategy because it's going to just get bundled into tax and not an accounting and practice management and workflow software. And if everyone becomes faster and more efficient, the tool-based competitive advantage is going to disappear. There's an advantage if you're an early adopter, right? But eventually it just like goes away. Here's the Spotify comparison. Spotify disrupted music two decades ago by giving enormous catalogs for a fixed price through inexpensive subscriptions or free access. And the artists benefited in the sense that they got instant global distribution. They didn't have to sign with a label. They could just go on Spotify. But every competitor did too, right? So having good music alone, it stopped being enough. Anyone could distribute their music. There wasn't this gate of the publishers. And so musicians then responded by building brands going on YouTube, getting online attention, interacting with fans, building up their own audience. And so Giles says that AI is going to do the same thing with accounting because the actual work product, the outputs are going to get cheaper to produce, easier to create. And so it's going to change what firms have to do to win. So what did musicians do? Musicians built their own identity. They created their own audience, not relying on a record label to have the audience. They published on SoundCloud. They got on YouTube. They gave their music away for free, built an audience, got emails. They sell LP's because they can make, you know, that more money selling one LP or physical record than they can with 50,000 plays on. Because it's a collector's item, right? It's a souvenir. Yeah. So I think building your own audience is like so critical now. I mean, it was important before, but now it's like that's a big differentiator right between you and any other firm that's using AI to do work and tech. So make a community, make your firm visible, create live experiences, bring people in to seminars, round tables, client events. That sort of thing makes a lot of sense to me. So basically, you know, the parallel is actually really good. It's that the thing that musicians sold, which was their music, because of online distribution, because of streaming, the value of that went to zero. This is a commodity. Yeah. Anyone can do it. There's no barriers. You can't sell it anymore. It's all free. Everyone's giving it away online. So you have to sell something else. And what musicians today sell is the relationship, right? They sell something that gives their listener a feeling of being connected to them, which is the t-shirt, which is the LP ticket. Yeah. So I think that's an amazing example. And anyone building a firm, right, should really be thinking about that. That's the thing that you build that technology is not going to be able to disrupt. So I have a story that actually proves that a little bit. And then I think you have one more music story, I think, about somebody in taxes, which you could follow up on this one. So you've talked about John Summon on the show before. I still don't really know who he is. He's a DJ guy, right? But he used to be the accountant. So listeners, it's not going to help for trying to get him booked on the show, David. You can't let's go listen to him on Spotify. But he, you know, so he used to be an accountant at EY and quit his job. He's a CPA crushing audit job. And he became a DJ and he's a CPA, right? And basically, like the number one DJ in the world, like one of the top DJs in the world. In the US, yeah. But I don't listen to his music because Spotify has 10 billion other choices, right? This is the issue, right? There's no radio. There's nobody pumping his music into my face. And that's not, I digress. But I've always said, right? That, like, and I tell my kids, like, getting a accounting degree because you know what career you have in the future, you will be successful because of your accounting degree. Well, some of it is actually crediting his accounting background for keeping him financially literate in an industry or artist routinely go bankrupt despite making millions. And so he's now taking some ownership models on music festivals. He's launching music festivals with 50,000 attendees sold out over two days. And he says he spends half his day on business operations. He said, you can't have acid. Now, if he was a chef and then he decided to be a DJ, he probably would not be as successful as he's currently successful. But because he has those skills, right? The accounting skills will carry you into any of their career you get. And he flat out credits it. There's no denying. Yes, he hated the job of accounting, but the accounting skills are making him a successful DJ. And the number one thing is, he's doing. You mentioned this is he's created his own record label. So now he owns the distribution from that standpoint. And he's he's booking other artists on it. And it's a festival. So he's having events under his own label called Experts Only. And they do the biggest one in New York City on Randall's Island. It's a two day festival. It just wrapped up. It was September 19th and 20th. And it attracted over well 30,000 to 50,000 fans over the weekend despite rain. And like that is an impressive thing. He now has that community and he's built a business around it. And so it's amazing, right? Like an accountant has become a global electronic music superstar. And now is building an empire. And the empire building is because it was a counting background. That's what's given the still set to do that. The discipline for sure, right? All right. Just the knowledge on how to do it. And I think you have one so you have an example that's the opposite, a musician that's the opposite of. All right. So John Summit, he was just in New York, a playing, you know, set for tens of thousands. Shakira, a global, a Latin music star Shakira is playing 12 shows in Madrid this month. And it's her first. She was not going to count, right? It has no accounting background. And these are her first concerts in Spain in eight years, which is a long time because I think she's really popular there. Well, why? Because she was in a 63, well, I'm not 63. She was 55 million euro fight with Spain. She won a tax court case there worth 55 million euros. And simultaneously, as she's doing these concerts, Spain's Supreme Court is reviewing her win. The government appealed. The whole dispute came down to one question. Did she actually live in Spain in the early 2010s? I don't know if she did. I mean, that's the question. It's all about residency, right? So I guess. I predict, you know, who won't have any tax issues? John Summit. His doctor's going to be in order. He's not going to have tax issues. I hope so. Yeah. I'm trying to think what came down. I want to see if I can see where the residency issue was. Yeah. I don't have the details here. Anyway, David, that's all the time we have for this week. Thank you to everyone who joined us live, tuned in on the podcast and YouTube wherever on Spotify. Please rate and review us those five star reviews, help get the word out about the show. And don't forget you can earn free continuing professional education for listening to this episode and many, many, many others and many other fine podcasts on the earmark app. Go to earmark.app in your web browser. We'll get the free app from the app store or Google Play Store. It's free to create an account, free to earn one CPE per week. And if you want to support the work that we do and get access to premium content, subscribe for the low price of $199.99 per year, just 200 bucks for unlimited CPE. We really appreciate all of our subscribers. And I was going to even better if you want to save 25% sign up for your team. Teams three, four, five people, you get 25% off. Sign up your team for your mark. You can do that right there in the app. Create a team account, invite your co-workers, get a discount. And if you want more discounts, your larger firm, email us. [email protected]. Let us know you're interested. We'll get you set up with a free trial. All right, David. See you around here next week. Thanks, everyone. Time for the Classifieds. Here's a chance to learn directly from the firms already leading the advisory shift and bring your whole team to do it. Advisory Amplified is back this fall, showing up in six cities, Rally, Nashville, Minneapolis, Boulder, Phoenix, and Seattle, September 22nd through October 9th. GA tickets are just $199. The whole day is CPE eligible. And this is an event built for firm leaders and their teams to support them in transitioning to more profitable advisory work. Go to advisoryamplified.com to find your city and get your ticket today. Can you defend your client's Shopify margin number? Profit root turns a Shopify orders export into a reconciled margin report. Every figure traces back to the source file, every coverage gap is named, and nothing is estimated to fill a hole. There's no sign up to try it. Walk the preloaded sample store or upload an export in the first report is free, built for accountants and fractional CFOs, try it at tryprofitroot.com. Want to get the word out about your newsletter, webinar, party, Facebook group, podcast, ebook, job posting or that fancy Excel macro you just created? Why not let the listeners of the Cloud Accounting podcast know by running a classified ad? Hit the show notes for the link to get more info.