Latest / The Accounting Podcast / Perplexity Goes After Tax Prep, IRS Testing Palantir for Audits

Transcript

Perplexity Goes After Tax Prep, IRS Testing Palantir for Audits · The Accounting Podcast

• Attention: This is a machine-generated transcript. As such, there may be spelling, grammar, and accuracy errors throughout. Thank you for your understanding! David Leary: [00:00:04] People have called Palantir the most dangerous company on earth. So that's who's going to be picking audit audits. Now for the IRS it's a little bit scary coming to you weekly from the OnPay Recording Studio. Blake Oliver: [00:00:20] Hey everyone, and welcome back to the accounting podcast, your weekly roundup of news in the profession. I'm Blake Oliver. David Leary: [00:00:26] And I'm David Leary. And Blake. I don't know with this being how many days out before the tax deadline here. Uh, six days, six days, seven days, six days. Yeah. If anybody's going to show up to the live stream today, everybody's probably pretty much heads down doing tax work. So we'll see. Give them a second. Blake Oliver: [00:00:42] We got lots of tax stories though. Stuff tax pros should be paying attention to. Perplexity one of the big AI companies out there. Maybe not as well known as OpenAI or anthropic, but definitely a big one. They're going after tax prep. You caught my attention with this David. We're going to lead with that and a story about how the IRS is using Palantir to decide who to audit, or they're scoping out doing that. They spent some money on it in 2025. And so maybe, you know, the company that is working with the Defense Department to pick out terrorist targets is also going to pick out audit targets. Um, and of course, we're going to talk about, uh, Trump accounts. They apparently they've been really successful. But before we do all of that and much more in the world of accounting news, David, let's thank our sponsors. David Leary: [00:01:28] Our sponsors this week we have on pay cloud accountant staffing reframe and worthy. Blake Oliver: [00:01:33] 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 is that? 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 Pei. Handles the heavy lifting. You get a dedicated onboarding coordinator who sets up worker profiles and transfers year to date data from previous providers, all at no extra cost. Their seamless QuickBooks and Xero integrations eliminate manual journal entries, and they support any type of business you serve farms, restaurants, nonprofits, you name it. On Pei can handle unique requirements without adding complexity and on. Pei keeps pricing simple, too. Everything your clients expect from multi-state filing to off cycle Pei runs is included. No hidden fees. No surprises. To book a demo, head over to The Accounting Podcast. That's The Accounting Podcast dot promo 1099 o n a y. Now, David, we got to get to perplexity and IRS audits, but let's do that Trump account story first because we talked about this on a past episode. And I'm curious to hear what's new. Like it's out, it's happening. Right? David Leary: [00:02:57] It's happening. And to rewind a little bit. For those of you who don't know at work, but basically any child that's born between January 1st, 2025 and December 31st, 2028 and has a valid Social Security number, is eligible to get one of these Trump accounts with the money in the savings account started for you with the money in the account. Blake Oliver: [00:03:17] And it's like not a small amount, right? It's like a thousand bucks or something. David Leary: [00:03:20] I think it's $1,000 from the government pilot contribution. But then some billionaires like the Dells, Michael Dell from Dell Computer, they're putting an extra $250 each to the first 25 million accounts that are on low income zip codes. But what's amazing about this is the totals that already the IRS says more than 4 million children have already been signed up for new accounts. Blake Oliver: [00:03:42] 4 million. David Leary: [00:03:43] 4 million. Um, and if you put this in perspective, and that's what I was trying to understand because, because really what this is, this is a government created savings vehicle. And the same way a 401 (K) is or a 529 College savings plan. So if you think about A 401 (K), about 60 to 65 million Americans participate. But there's 160 million workers in this country. So it's maybe 35, 40% of the people participate in 401 (K). And that's including employers who a lot of them auto enroll you. Right. And then you think about 529 college savings plans. There's about 15 million, 16 million of those, but there's 70 million children. So again, that's maybe 20, 25% of families. And if you think about the time for one K plans, it's been decades and they're only at 3,540% participation. College savings plans. It's been 25 years, 25% participation. It's been about a year. And they're at there's 4 million accounts for about four and a half to 4.7 million kids. So it's pushing 85, 90% adoption in one year, which is this is really amazingly successful. And I think the reason why is because it's just a form. You just fill out that whatever the stupid name is because Trump put his, uh, is it 45, 47? Blake Oliver: [00:04:57] Yes. Form 4547 Trump account elections. David Leary: [00:05:01] But it's a one page form and you just file it with your tax return. So the barrier to entry is, is relatively a lot smoother than some of these other like I don't even know how to set up that chill. The, the savings plan for the college. Like that's way too complicated and you wouldn't do the 41K if the employer didn't offer it. But this is I think it's because they've made it simple for people to enroll. And then the other carrots there, there's money, right. So it's a combination of both. But but but to have 80%, 90% participation is for anything the government does is very. These are crazy successful thus far. Now, it doesn't mean that these funds will get locked in with other legislation one day. And then these kids can't get their money. Who knows? But as of right now, it is a successful government program. Blake Oliver: [00:05:49] And I got to say, the website actually looks pretty good. Trump accounts.gov The headline is Trump Accounts Jumpstart the American Dream. These are tax advantaged investment accounts for US citizens under the age of 18. Fill out IRS form four five, four seven when you file your taxes, or by clicking below to make your election for the 2026 tax year. So this is all children under 18. So when you did that calculation, David, on the success, like you were dividing the number of accounts that have been created by what population? I'm just wondering how you figured that out. David Leary: [00:06:25] For the Trump accounts. Blake Oliver: [00:06:26] Yeah. David Leary: [00:06:27] Um, it's the number of births that have taken place since those two days. Blake Oliver: [00:06:30] Oh, well. But can't anyone. David Leary: [00:06:32] January 1st of 2025. Blake Oliver: [00:06:33] Can't anyone get these if they're under 18? David Leary: [00:06:37] No, but you have to be born in that window. Blake Oliver: [00:06:39] Oh. Got it, I see. David Leary: [00:06:41] Which is by default. You'd have to be under 18. Yep. Blake Oliver: [00:06:46] Well, that's, uh, that's interesting. You can contribute. Families can contribute up to $5,000 a year. And according to this math on this website, if you did that for your kid by the age of 18, they would have $271,000. That's based on an account opening at birth with a $1,000 opening deposit. And the rate of return is derived from historical S&P 500 averages. That'd be pretty nice. Turn 18 and you get $271,000. Maybe that would be a down payment on a house someday. All right. Welcome to our live stream, viewers. Hey, Mesay doctor or Dre the dream, Edgar Decker, CPA. Anis. Oh, Heather Smith. Hey. Hello. Great to see you. Welcome, everyone. Don't forget, you can earn free continuing professional education for listening to this episode and past episodes of the Accounting Podcast and many other fine accounting, tax and audit podcasts on the earmark app, go to earmark app and your web browser, or download the free app for iOS or Android. Create your free account. Earn one free CPE every week. And if you have an accounting firm, we offer team subscriptions. Contact sales at. Earmark CPE. Com to get set up with a team account for everyone in your firm. We offer great discounts for firms that sign up for more than. What is it? More than ten users. Five seats. Oh, more than five seats. David Leary: [00:08:11] And it just goes up from there. Blake Oliver: [00:08:12] So if you've got more than five CPAs and EA's at your firm, get in touch with us. Sales@earmark.com. We also have this amazing feature called a private channel, where we can take your lunch and learns your internal trainings, and we can turn those into self-study CPE that your team can get anytime, anywhere. And it's an awesome feature. Nobody else has it. Get in touch if you want to learn more. Okay, let's talk about perplexity. They're going after tax. You spotted an ad in your Instagram feed, right David. David Leary: [00:08:45] Yeah. So I'm going to share this. This might be Instagram or Facebook. So I'm sharing my screen. For those of you that are just listening only, it's a simple black and white ad and all it does is it says, find every duplicate duplicate entry hiding in your QuickBooks. Blake Oliver: [00:08:59] So I said text, but this is actually for like bookkeeping. David Leary: [00:09:02] What put me on the tax discovery. So, so I saw this. Of course, I'm going to click through. I'm a sucker for these things. I'm like, why is perplexity, you know, advertising about QuickBooks. And it reminded me of the ad I saw two weeks ago when I said, I'm starting to get targeted ads for accounting workflows from Claude. Remember two weeks ago, we talked about this two episodes ago. So I click through and it takes me to a web page and it's called audit Your QuickBooks financials. And they refer to this as their it's a health check. And so they can do things like PNL analysis, um. Blake Oliver: [00:09:38] Analysis. David Leary: [00:09:39] Analysis, sorry, Typekit detection, expense categorization, AR and AP aging reviews, reconciliation audit and a financial health score. So this is arguably this is all the stuff that people have been building and doing with AI in their QuickBooks file, right? For the last, you know, two years. Blake Oliver: [00:09:59] Yeah. David Leary: [00:09:59] For experimenting with AI and or third party apps are building this functionality now. Perplexity. They launched a new thing called computer, which is basically their version of like a cloud coworker or these agents, right. And so then I was like, man, I haven't been to perplexities website in a while. So I was like, let's just go to perplexity dot AI. I click over there and this and it looks a lot like Claud. It's got that, that tan color, but it actually has like suggest, you know, like Google has suggested searches. It has like buttons and suggested, uh, instructions. Yeah. And it has the very first one it shows was navigate my taxes and then it says, prepare my 2025 taxes for review, review my prepared tax return for errors. Uh, share in the questions specific to tax. Right. So I saw that and I'm like, oh, perplexity is going right after core workflows QuickBooks and taxes. Then I came across the news article later on today when I was prepping for the show. Yes, they've officially launched what they're calling is computer for taxes. It's an AI agent that goes beyond answering questions to actually drafting full federal tax returns. So it's shifting from AI as an advisor to AI as an executor. The tool reviews, uploaded documents, ask follow up questions, and maps everything directly onto IRS forms. And unlike traditional chatbox, it continuously updates the tax modules tied to the current IRS rules. So if you remember before when we first saw these previews of people trying to like upload tax information and then get it to do their tax return, it was one of the first demos of ChatGPT, I think remember back three years ago. Yep. Those are the early demos. But then the issues were the model was never trained up to up to date on tax stuff. And then you have all these startups show up. Even though you talked about who was it last week? Tax GPT, right? Blake Oliver: [00:11:49] Tax. Gpt says that they have automated 1040 tax prep from tax. David Leary: [00:11:55] And then they're uploading and maintaining the IRS information. Well now out of the box perplexity is just doing that. Blake Oliver: [00:12:01] I'm going to try doing my tax return with Claude Cowork. And if I can get my hands on a Mac mini, maybe I'll install Perplexity computer on it and see if it can do it too. I guess I don't need a Mac mini to do it. I can just do it on my on my on my laptop here. David Leary: [00:12:18] But I have not played with perplexity computer, but I saw another ad from perplexity. Now that you go to the website once and it looks like you just do this in your browser, there is no installing and doing other stuff. That's why it's a little bit safer from a security risk. Blake Oliver: [00:12:32] Well, so what's a big deal about Perplexity Computer is that it can interact with the desktop apps on your computer, not just what's in your browser, not just what's in the chat window. And this is where all these AI agent tools are going. Claude Cowork has that feature in beta. So Cowork just came out like today or yesterday as a publicly available tool. Now it was in some sort of beta. Before that, I had access to. So now anyone can use it. They've also got the desktop interaction. And what makes these AI agents powerful is that they will first try to use an API or in AI tech speak the MCP. They will try to use the MCP to talk directly to whatever app you're working with. If that's not available, it will open up a Chrome tab and then interact with the app in the tab. And if that's not available, then it can try and use your desktop. So it's going to have all the different ways that we interact with apps. And it's slow now. Like it's still slow because every time it clicks as take a screenshot, interpret that screenshot, and then decide what to do next. So it's kind of infuriating sometimes, but it's getting really reliable. And I have a story about using co work on some bookkeeping work for our company. David Leary: [00:13:51] I had a couple more thoughts on this as I'm observing these AI models kind of role in this functionality is part of their base functionality. And if you think about QuickBooks and zero, they keep talking about how the data is remote. The data is the moat. Well, if these models are connecting to the QuickBooks data or zero data, learning the data, maybe they're not copying the data, they're just learning it. What's next? It'll be like spin up your own QuickBooks, because all it takes is one of these AI companies to either partner with a company like Supabase or a database company, and it just spins up a database, creates all the tables that basically are what are, which is the fundamental basics of QuickBooks or zero. And use your UI, just build your own. Like this is a. Blake Oliver: [00:14:36] They're already doing it. They're already doing it with Excel. So Excel is like the default database now for cloud Co-work. If you want it to organize and structure data, you can tell it to build an Excel sheet that does that. A whole workbook and I did this also. This is one of my examples is I've been tracking a few deals that we've been working on. Right? Firms that want earmark for teams. And I was just tracking this in a very, you know, loose way, right? Like some files in Google Drive, right? Nothing structured. And so I decided to bring that into cloud Co-work and I had it create a CRM in Excel. And it is now going through and reading my emails and my calendar and my meeting notes that are in fireflies and in Google Drive that Gemini captures. And it's updating the spreadsheet for me. And the spreadsheet does stuff like calculate the pipeline value, the weighted pipeline value, the number of active deals, the average deal size probability. And like this is something that a sales manager, his entire job would be. And you know, I haven't really dug in to see just how accurate it is, but it looks right. Like it looks directionally correct. David Leary: [00:15:59] And, and that's the thing, like it's good enough. And we'll talk about this in a second story. I have about digits. We'll talk about good enough. But when seeing how Claude last week or two weeks ago and now seeing this with perplexity are these startups like accrual basis tax, GPT, are they kind of hosed? Are they going to be dead? I mean, they're starting they're building something similar that's proprietary. They're starting a brand new company. They're getting lots of investment money. And the fact as soon as they're I mean, ChatGPT just launched a week ago and Claude or not, Claude, whatever this is, perplexity is doing what a whole company just launched, a whole business launched to do what is now a stupid feature in perplexity. Blake Oliver: [00:16:40] Well, whether or not there are like the question is, are these just wrappers right? For this technology that anthropic and perplexity and OpenAI have built. And I think the answer is basically, yes, they're they're using this tech and they're creating an interface for it and adding value. And I think to answer your question, David, we have to think about what is it that they're adding that these, let's call them AI platform companies or MLM companies are not doing. And I see it as two things with accounting firms. One is auditability. Yeah. The accounting firm needs to have agents working in a way that's trackable, traceable, defensible, all of that. Um, another way is limiting the sources. So making sure that these agents are in fact only relying on the right information. That's what tools like Bluejay are doing with tax research there. They're using this dedicated database that the agents are using, not just anything. And who knows what the perplexity team actually uses the database for their tax tool. Right. I mean, that matters for accuracy because there's so much conflicting information and outdated information out there. Um, another one would be just being able to deploy these at scale, because a lot of these tools right now are just built for individuals. And if you want to implement this at your firm, you've got to have a system that everyone can use the same way to deliver a consistent output, right? If you've got a bunch of different preparers and reviewers doing tax returns, they need to be doing them the same way. David Leary: [00:18:22] And then how is the. I imagine it's just creating a PDF of the tax form filled in. So some of these proprietary platforms, these newer ones like filed, etc. I think tax GPT does it too. They pass the data into Intuit Proconnect. Yes. Blake Oliver: [00:18:38] Yes, they actually go. David Leary: [00:18:39] Enter that are out there. Blake Oliver: [00:18:40] They actually go enter the numbers into the forms into the tax software. David Leary: [00:18:43] And so that way you can push it to file it, etc.. Blake Oliver: [00:18:46] Push it to you for review so you can do all your work, right. They, they integrate with your current workflow. David Leary: [00:18:50] But is that just a matter of time? Blake Oliver: [00:18:52] I mean, that's. David Leary: [00:18:53] Six months away from this now. Blake Oliver: [00:18:55] I don't, I don't know, but, but see, the thing is some of this stuff you'd have to like consciously build and I don't know if like perplexity is going to like go think about accounting firm specifically. They're building these general purpose tools. David Leary: [00:19:04] So you're totally thinking about accounting firms. Why am I getting targeted ads that are obviously targeting me thinking I'm an accountant? Know that very targeted. Blake Oliver: [00:19:13] Isn't that ad targeting you as a business owner? It's asking you to check out your QuickBooks health. It's doing a. David Leary: [00:19:21] Quickbooks health might be true. Okay, I'll give you that. This ad might be, but. Blake Oliver: [00:19:25] That's who they're going after. They're going after small business owners who would normally have to hire an accountant or a bookkeeper, a CPA, and they're going to build tools for that end user. David Leary: [00:19:37] Or get a third party app. I would hate to be a third party app developer right now building for the QuickBooks or Xero ecosystem. Blake Oliver: [00:19:43] Well, it's tough because yeah, a lot like like, for instance, all the receipt and bill entry apps, they're, they're toast already. David Leary: [00:19:50] And the one nice thing about us doing this show now for 8 or 9 years going on nine years is I have this search, all these searches I do and I haven't really changed them much in eight years. And so, but the stories that are in the results of the search have changed over time. And every week now I see a dozen, like either Reddit posts or other posts where people are like, I built this, I built this, and they're like small business apps. People are just building, they're building their own gels. They're building their own receipts. Captcha thing, they're building inventory apps. They're just people are just building their own stuff at a pace that I've never seen before. And I didn't, I'm not searching for I've always had just search terms like small business app, right? Quickbooks zero. And now, now these things are bubbling up. I'm not going my way to find these articles. They're just more of them. Blake Oliver: [00:20:39] All right. David Leary: [00:20:40] So perplexity but yeah, they're coming. They're coming. Blake Oliver: [00:20:44] Well, okay. Yeah. I mean, so here's my view. My take is that solo practitioners, small firms where it's, you know, one partner can really decide the tech stack, you're probably going to be able to use cloud Co-work or perplexity computer or whatever this tool is and customize it to your liking and use it. And, you know, if you're tech savvy, you can do that and you don't necessarily need a specialized AI agent tool. But if you're a big firm, I think there's an opportunity there. That's why basis raised so much money. That's the opportunity. That's why tax GPT is doing well. Like they're going to create standardized workflows around these agents. David Leary: [00:21:24] That's my or if you're a big firm, you could just build your own standardized workflow using these. No code like this. We're in a funky time. Yeah. It's crazy. Blake Oliver: [00:21:32] It's exciting. I was just talking with, um, uh, a few different people about this, and I feel like we're going to be able to do the work of ten people in a firm. Like a ten person team will be three, and it'll be like a partner and a, like a manager, and then maybe like an admin or tech person and. David Leary: [00:21:51] Or a salesperson. Blake Oliver: [00:21:52] Yeah. Well, in a small firm, it's going to, you know, it's going to be you the partner, right? So, so, but like three people could do the work of ten. And I think that's possible, like starting now. So just imagine. And that doesn't mean you actually cut people. It means that maybe you can take on three times as much revenue with the same team. And so the opportunity for small firms to like increase profits, to really make good money and be competitive as like jobs in that sense, like is really. David Leary: [00:22:24] Instead of having 30 clients at your firm, you now can have 90. And that makes a huge difference. Blake Oliver: [00:22:29] Um, All right. Let's talk about the IRS using Palantir. But first, let's thank our next sponsor. And that is cloud account and staffing. Are you tired of the endless search for qualified accounting talent? You are 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, 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 ten minutes. Head over to The Accounting Podcast. That's Accounting Today dot promo c a s. Oh, and I forgot to talk about what I did with cowork. David Leary: [00:23:46] Yeah, talk about that. And then I'll talk about digits before we jump to the IRS. Lets. So tell me what you did with coworking. Blake Oliver: [00:23:53] So I've been seeing some like influencer reels on Instagram talking about how they're using AI for bookkeeping, like specifically Kofiwork. So I decided to give it a try. And I've been, I've been learning about Cowork over the last week. You know, setting up projects, that sort of thing. And I created a bookkeeping project and I wanted to get it to help me close the books, reconcile my bank accounts. I figured that would be a good starting point. And I was able to successfully do it. And not just I didn't have like a ton of transactions because I wasn't going to try this on like a big account, but it's like, um, I tested it on a brokerage account. So this is an account that we have and it's got interest. It's got transfers from different bank accounts. Not a lot, but it's also not connected. So there's no zero connection. So we've got we've got some transactions posted to the GL but no bank feed. So what needs to happen is I guess what I was doing in previous months is I would just go in and look at the account transactions, enter any that were missing and then manually reconcile them. Okay, so I figured this would be like a good task for AI I'd give it. David Leary: [00:25:07] Because there's no there's no half reconcile because there's no bank feed. So nothing's getting half matched or you're not getting that, that jump start on the reconciliation. Blake Oliver: [00:25:14] So I started up a chat in my bookkeeping project in Cowork and I said, here's the bank statement. Go find this account in zero and reconcile it. And I didn't have to give it a lot of instruction because that's the beautiful thing about Cowork. It will create like a task list, and if it doesn't know what to do, if it needs more context, it will ask you questions. So start asking me questions about what to do. And it surprised me because it didn't just go and take the bank statement and then enter in missing transactions. It actually took the bank statement, parsed it, turned it into a CSV file, formatted the file like zero, needs it to import as a bank feed. Imported the file as the bank feed the bank statement. Then it matched the transactions to the GL the same way I would do it. Having to go in and click okay, okay, okay. And it did that in the in the Chrome browser. And then it ran the reconciliation report and saved it and exported it as a PDF. So it did full end to end And it asked me questions, a few questions along the way. And then here's the really amazing part, is that once that was done, I asked it to save this as a skill. There's a, a skill in clod called skill creator and skills are basically just like saved instruction sets. Saved prompts, if you will. David Leary: [00:26:46] Yeah. Which, which, which then they save and then they're going to all of a sudden start putting that out there. Blake Oliver: [00:26:51] So those, those like those prompts that you've seen as suggested ones in perplexity, it's the same idea, right? It's these prompts that they've iterated on and they think are good enough to share with the world. And what's amazing is that it saved this skill. And now I can reuse that. So I can basically upload a PDF and I can say reconcile this account and it will do everything as I like it and it will then offer to improve. David Leary: [00:27:18] You have to report back. Back. Yeah. Blake Oliver: [00:27:21] And it will then improve the skill. So if like the skill doesn't do what I wanted, I can give it feedback and it can go and edit the instructions on its own and improve the skill. And the instructions are super detailed, way better than anything that I would write. And it learns somehow, like goes and figures out what to do. And if it doesn't work, it saves it in its memory and that skill. So it doesn't make the same mistake again. So it goes faster every time. So basically what I'm trying to say is that right now, in a simple scenario, you can end to end reconcile bank accounts in small business software. And I don't see why this wouldn't work with more complex accounts with more transactions. David Leary: [00:28:01] So this makes me question like, what is the point of jaks in zero then just. Blake Oliver: [00:28:06] Zero. David Leary: [00:28:07] Ai. Blake Oliver: [00:28:07] Data inside of it? Well, I mean, that's why do. David Leary: [00:28:10] You have to build this, right? Blake Oliver: [00:28:12] I don't think they do. I don't think they do. David Leary: [00:28:14] Now why do you have to build it? Does that make any sense? Blake Oliver: [00:28:18] Well, it's because it's because when zero started building jacks, they built it, copying the chat bot functionality of ChatGPT. And that's limited in the sense that you're limited to that conversation. It doesn't know how to do extended thinking, it doesn't create this checklist, it doesn't follow the. David Leary: [00:28:36] Browser plugin, etc., etc.. Blake Oliver: [00:28:37] And so now all these AI companies are just racing ahead of everyone trying to build it themselves in their products. And so zero made the right decision. Like we talked about last episode, they're just integrating with cloud. So I'll be able to in clod. Clod will not have to go into the browser to do this stuff. It'll just do it via the MCP, which is basically the AI version of an API. And it will talk to zero and do all this stuff not through the interface, the user interface, and same thing. Eventually with QuickBooks, when QuickBooks customers demand it and Intuit has to do it because their own AI is just not going to be able to keep up. David Leary: [00:29:17] Yeah, I mean, they have the one now where it'll scan the bank statement for you and then do the matching. But it only they kind of are cheating because they only let you do that so far with accounts that have bank feeds already. So it's half cheating, right? It's less impressive. So the counter, it's already been matched. So it's only how hard is it for the AI to do the last 1 or 2, three transactions. Blake Oliver: [00:29:42] So I know that all of this is it's like a little bit difficult to hear if you haven't been using these tools or even if you have, because so far most AI implementations are not saving a lot of time. Or I should say this, they're not delivering like business value. And I saw a story about that that I want to talk about. I'll bring up some stats from that where most organizations that are implementing AI are not AI are not seeing, they're not seeing super big productivity bumps. And I have I know why that is. And I know why in accounting firms, we're going to struggle. Um, but before that, let's talk about the IRS using Palantir. David Leary: [00:30:25] I'm going to push that one out again on you. I'm sorry. Why? Because digits is related to this AI stuff. Before we jump off to the other. Blake Oliver: [00:30:32] Okay, okay. David Leary: [00:30:33] So if you think about the world we lived in before we had SaaS and we'd pay for SaaS. How would we pay for that, Blake? 30 bucks a month. 60 bucks a month, right? Blake Oliver: [00:30:42] Sure. David Leary: [00:30:43] For SaaS software. But now AI has came along. And how are a lot of these AI models charging? They're charging for usage or tokens. Right? Well, digits just announced and I think it's really, really smart. They are now going to switch to outcome based pricing. Blake Oliver: [00:30:57] I saw this. David Leary: [00:30:58] So instead of charging firms for access, it's going to charge when the platform actually automates enough work to matter. Um and it's only going to kick in billing when the client reaches 95% of the transactions or zero touch before the month's close. So if you think about this, this is really like a it's a trial. It's a win win, right? So you now get to try digits, but never pay for it until digits actually delivers and earns their money well. Blake Oliver: [00:31:25] But here's the catch. David Leary: [00:31:27] What's the catch? Blake Oliver: [00:31:28] Well, think about it. Let's say they automatically categorize 95% of the transactions. Does that mean they categorize them correctly? David Leary: [00:31:36] Huh. Well, if you have to fix it that means you touched it, right? So the way they're defining zero touch transactions, transactions that are not created or edited by a human accountant before the books are closed. So if it miscategorized something, okay, edit it. That would not count as the 95%. Blake Oliver: [00:31:53] Well, it sounds like it's like a guarantee. That's kind of nice. David Leary: [00:31:56] And so what's interesting about this is I think this is going to allow digits. Right now. Digits is $100 a month for their middle plan. Yeah. This is going to let them charge 300 $400 a month. I could even argue they might be able to charge 500 bucks a month for it, because at that price, it's $6,000 a year to get 95% of your bookkeeping done. Blake Oliver: [00:32:15] Right. David Leary: [00:32:15] Nobody's going to spend the extra 50 K to have an in-house bookkeeper for the extra 5%. Blake Oliver: [00:32:20] But here's where digits, I think, is going to have a hard time, right? This is here's the challenge for a company like digits is I can use AI agents to just do all of that automatically in QuickBooks or zero. So why am I going to switch. David Leary: [00:32:38] My switch to GL? Yeah. Blake Oliver: [00:32:39] Why switch the GL, which is really hard. It's a lot of work to switch to GL. So I think that's going to be profit. David Leary: [00:32:45] Loss and balance sheet. That's exactly like the one you have today. Blake Oliver: [00:32:48] Exactly. So you know if these MCP servers open up with our core deals with QuickBooks, then I think it'll help Intuit just stay dominant forever. But I mean, that's part of the problem is that like Intuit doesn't have like API access to everything. And really you need API access or MCP access to everything in order to do this stuff. All right, let's thank our next sponsor. And that's the reframe conference. Are you still billing by the hour? Still stuck in the compliance grind. What if I told you there's a way to transform your accounting practice into a thriving advisory powerhouse? That's exactly what reframe 2026 is all about. This November 5th through seventh in Miami, the accounting profession's most innovative minds are gathering to flip the script on traditional practice. Co-hosted by Carlos Garcia and Hector Garcia, CPA, reframe brings together industry legends like Ron Baker, Blair Ends and Chris Doe to show you how to move beyond transactional work and into high value advisory services. You'll discover how to package your expertise, lead powerful client conversations, and build scalable models that drive real impact. This isn't just another conference. It's where accountants come to revolutionize their approach, automate their workflows, and humanize their relationships. For four years running, reframe has sold out because forward thinking professionals know this is where transformation happens. To get your ticket now before they sell out again for the fourth year in a row. Head over to The Accounting Podcast dot promo slash reframe 2026. That's The Accounting Podcast dot promo forward slash reframe 2026. Scott in the livestream says it's those AI spam bots sabotaging Blake since he let them onto his personal computer. David Leary: [00:34:35] Yeah, it might be. Blake Oliver: [00:34:37] Scott also is very skeptical of, uh, of AI. Where's his comment above? He said, you trust these programs that much with your company's data? Bit of a security risk, I know, but hey, no risk, no reward. I live on the edge, Scott. I just like to mountain bike. David Leary: [00:34:59] I'm really worried about your data. Let's talk about Palantir. Blake Oliver: [00:35:02] Yeah, yeah. Yeah, exactly. You're under Scott. You should be way more concerned about the US government using AI than than I should be concerned about, my company's data. Okay, David, you got this story. David Leary: [00:35:16] Yeah. So the IRS is piloting a new AI powered tool built by Palantir to modernize. Modernize how it selects audit targets, aiming to identify the highest value fraud and noncompliance cases more efficiently, efficiently. They're calling this system snap, which is interesting because isn't snap like the. Blake Oliver: [00:35:34] Uh, food stamps. David Leary: [00:35:35] Food, food stamps thing. It's kind of interesting. It's given the same name, which is weird. So this is going to sit on top of the IRS fragmented legacy systems and analyze large volumes of structured and unstructured data to surface red flags that human auditors might miss. So they're going to pay them 1.8 million to develop this tool. The pilot and Palantir already receives $200 million in IRS contracts since 2014. And so what this means, if you think about the old ways, the old rules of thumb you might have had on whether or not a client was going to get audited, that's probably going to change. Blake Oliver: [00:36:11] Yeah, definitely. David Leary: [00:36:12] Because they're going to look at different data and the, the old buckets of the way they used to do it is not going to be there anymore. Blake Oliver: [00:36:17] And that's because David, they used to look at the numbers that went in the tax return. And so if those numbers didn't line up with benchmarks, they'd flag you. David Leary: [00:36:29] Yeah. There's like a tolerance. Blake Oliver: [00:36:30] What Palantir is going to do is look at the unstructured data. And so we don't know exactly what that is. But you know, it might be like disclosure materials that you include with your return that are not actually part of the data that gets entered into that IRS system. The accompanying information, like support or. David Leary: [00:36:51] Or maybe it'll be data that's at the Department of Defense. Has you or the Department of Homeland Security has on you or law enforcement has on you, because Palantir is offering services for all of those government agencies as well. So think about it. Pentagon, CIA, Ice. Now, IRS people have called Palantir the most dangerous company on earth. So that's who's going to be picking audit audits. Now for the IRS it's a little bit scary. Blake Oliver: [00:37:18] And they could be comparing with the IRS has to anything that's publicly available, like you said. And one of the examples in this story from wired that I saw, um, is like your e-commerce storefront, like your Etsy or your Depop. And they could look at that and then see what you are reporting and make some sort of calculation as to whether or not you're potentially underreporting your income based on what you're selling or what you're posting on social media. David Leary: [00:37:51] This is going to go sideways and probably abused, uh, by. Blake Oliver: [00:37:57] I don't know, I mean. David Leary: [00:37:58] Executive. Blake Oliver: [00:37:58] Branches. Look, I'm, I'm all in favor of like better targeted audits. I think every accountant would agree that we need audits to keep people honest. If you don't audit, then people will. Too many people will cheat on their taxes. So you got to have audits. But we don't want to audit the people who are honest. That just wastes their time. David Leary: [00:38:23] If you use. Our government has separate government branches for a reason, right? You. The checks and balances are there. Sure. Having the same AI company do the law enforcement and IRS enforcement is a little scary. And military. It's deciding where we're going to drop bombs. Blake Oliver: [00:38:38] Oh yeah, I know like the big brother implications are a little bit frightening here, but I don't know, it's I've sort of just resigned myself to the fact that in a digital world where everything's online and we've got AI running around, privacy is going to be something that's harder and harder to come by. What can we do? David Leary: [00:38:58] That's why you should always like post check ins at a Red Lobster in Alaska every so often, just to throw off the confusion. Confuse the algorithms. Blake Oliver: [00:39:06] David's going off the grid. All right. Let's talk about that. That story I teased about companies investing a lot in AI, but not really seeing much in the way of results. And there's actually an accounting angle to it. So this is from CFO comm. The story is your business. Your AI business case is built on the wrong math. It's an opinion piece by Alexander Hilton. And he argues that companies are failing to see meaningful returns from AI because they're evaluating it with the wrong accounting method. They're using cost accounting, which rewards local efficiency gains rather than throughput. Accounting, which asks whether AI removes the system's main bottleneck and improves financial performance. 80% of firms reported no measurable AI impact on employment or productivity over the last three years, despite all the billions and billions of dollars that are getting put into AI. And if you work at a firm, you probably have experienced this. So why is that? Even though those of us who use it can see the productivity gains and the author Hilton says that the gap between spending and results resembles Solow's paradox. Technology is visible everywhere, but productivity gains are hard to find. And it's the way we're looking at AI that's causing the issue here. So this is this is what goes to cost accounting. And I think cost accounting, honestly, is like the root of all evil in the accounting profession at this point. David Leary: [00:40:58] There's a whole book about this, right? I think it's even called. Blake Oliver: [00:41:00] Cost accounting was an incredible invention that changed the world in the industrial era. But we keep using it in in a world of intangibles and it's hurting us. So here's the example I use. Cases will focus often on hours saved headcount reduction, or just like speeding up certain processes, which I guess is also like hours saved, right? Sort of, not always just speeding up like a local process, like a smaller one. And the problem with that approach, when you're doing AI, if you're only looking at like hours saved, is that it assumes that all these local efficiency improvements, all these small process improvements are going to help the business overall, but they often don't because there's another bottleneck somewhere down the chain. So an example of that would be in accounting automating the tax return prep with AI. But all the returns pile up, even though they're done almost instantly, because there's a review step that has to happen. And so the returns don't actually get delivered faster to the client. You may have theoretically saved all these hours on the prep, but it's not getting to the client any faster. You're not delivering any additional value to them. And maybe you haven't even saved any headcount because the hours that you've reduced, you know, didn't actually take away somebody's job. That person still has to take all the files and put them into the AI to get it to generate the return. Right. David Leary: [00:42:40] So because you accelerated some part of the process over here, but the bottleneck. Blake Oliver: [00:42:44] Exactly. David Leary: [00:42:45] There's a blocker down downstream. Blake Oliver: [00:42:47] So he says that what we should be measuring instead is throughput. It's the rate at which the organization generates money through sales. Mine is truly variable costs. And in a professional services firm, hours are not a variable cost in general because most employees are salaried, so their hours are just an imaginary cost accounting construct. You're just dividing the the time that they are working through the day into these units. And yeah, you might save those, but saving those doesn't actually save money. And saving time on an individual tax return prep doesn't necessarily increase the amount of money you can charge for the return or speed up how quickly you can deliver it. David Leary: [00:43:31] What matters is are you doing more returns overall? Blake Oliver: [00:43:34] Yeah. Can you? But the problem is if the review is the bottleneck, you can't. Right? You can. You can increase the you can reduce the time to do the prep, but you still got to review all the returns and you don't have more reviewers. You can't do more returns, so you don't make any more money. David Leary: [00:43:51] This is what you've been saying for a while. This this like firms with bad processes that are going to try to implement AI are not going to be saved by the AI. Blake Oliver: [00:43:59] Right? David Leary: [00:44:00] If processes are, it's going to make him make him seem worse. Blake Oliver: [00:44:02] Actually, if you speed. David Leary: [00:44:03] Up. Blake Oliver: [00:44:04] A bad process or like a. Yeah, speeding up a process doesn't necessarily make it better. Might make it worse, right? Um, so that's, that's the issue. Yeah. It's, it's like increasing the speed at the beginning doesn't necessarily help you at the end. So I thought that was a really interesting way to look at AI. And so what instead we should be doing is looking for the current bottleneck and applying a out of that. And if you haven't found the bottleneck, it means you're not looking in the right place. So you have to look at your, your current process and then say, what is taking us the time? What is keeping these returns from getting delivered? Or why is this audit not getting done? And look for that spot and then automate that and then you'll actually see improvements. Aw. And of course, increasing the value of what you're delivering to the client so you can charge more. David Leary: [00:44:58] And by human nature, people are not self-reflective. So people are going to be like, obviously it's the client intake form, or I have to data entry, the forms faster or whatever it would be. And it's never going to be the reviewer person in the process or the senior partner signing off. Yeah. Blake Oliver: [00:45:11] And that's totally right. It's, it's wait, what did you say, David? You said the. David Leary: [00:45:16] The senior partner, the reviewer signing off. That's going to be self-reflective, right? You're always going to this is why oh, we need a better client intake form that has AI. So I'm going to go implement that. But then you never see what the spend back because you're the actual part that needs to be automated. Blake Oliver: [00:45:29] Well, the actual part in my experience, that is like the biggest delay in almost any accounting process is getting the source documents, getting the information that you need to do the return, the actual like prep, while it might take hours, is not nearly as, as much of a delay as just getting what you need from the client. So that's where we should be applying. Ai is figuring out how to You automate that with AI agents? You know. David Leary: [00:46:01] Let's say you do that and now you're getting the intakes faster. The reviewers, you still have this bottleneck over here. And that's what I'm saying is like, people have to step back and like, where is the true bottleneck? Because if you solve the one good. Blake Oliver: [00:46:14] I got a comment here. So I think this is Joffrey. Joffrey says to me, the bottlenecks are one client incorrectly enters data. That's definitely I mean, that's, that's even that's, that's a bottleneck because it just sends you kind of back to the beginning, right? Confirming, let's say, let's say the bottleneck is like confirming the client data to blurry forms. So the client sent you something you can't actually read. Now you got to go back and ask them. David Leary: [00:46:41] Maybe on purpose. Blake Oliver: [00:46:42] Like they take a picture with their phone and they send you an image via a text that's like too small to read. Um, incomplete forms. They don't give you the data in the first place. Multiple forms is one PDF having to go then separate all that out and then five invalid file format. So all of this is related to just getting the data that you need to do the return. So I can see a few different ways to automate this, which is have an AI agent that monitors a client folder and then compares that to a list of documents that we've requested and in a human like way, follows up with the client to get the missing information. But that takes a lot of human touch. Knowing how to do it without harassing the client, making them feel like they're getting bugged. And also clients will ignore automated stuff. But I think that's a real possible like point to solve a pain in for AI companies. I'm curious to hear if any of our listeners in tax or in audit or in just financial accounting or bookkeeping have other bottlenecks that they experience in their current workflows. Maybe we can brainstorm how AI could be applied to that. I'm just thinking right now in cloud Cowork, I could have a, a project called like, maybe I'd have a project per client or I could have a project just for tax prep overall for my firm. Blake Oliver: [00:48:16] But either way, I could, let's say it's my client that keeps it easier. I could use Kofiwork to create the the PBC list, right? The, the document request list, basically the help, you know, the client know what to send me. So give them give the agent last year's tax return. Give it everything that I know about the client. Just dump that all into the folder and say, create the request list for the client and make this as an Excel document. And cloud will go review the return. It'll go review all the source documents and put together a nice little excel sheet of everything that we need. And then I could say to Co-work, send this along with an email to my client asking them for all this stuff. And then I could potentially I'd have, we'd have to figure out how to do it. Right. But you could like, maybe you have like a client portal folder. And this is where the, the project is working from. When the client uploads documents, they go into that folder. And you could schedule a task to run every day that reviews what the client has submitted, updates that tracker, and then sends an updated list of requests to the client with whatever is missing, whatever was illegible, right? That admin type work. David Leary: [00:49:40] And and that's the real like I saw an article from the Sastre guys again and they talk about it's consistency. Like you're getting consistency from AI. Ai is not going to be tired on day two or day three or day four every day. It can go review these things again and send the email and ask for it again and do it again tomorrow. And do it again tomorrow. And do it again tomorrow until it doesn't have to do it anymore. And it's never going to complain. I don't feel like doing it again. It's not going to have like, I didn't have enough coffee and I missed some. Then tomorrow I saw them. In theory, you're getting that consistency over and over and over and over again. Yeah, 24 hours a day. Blake Oliver: [00:50:14] So that would be something I would try building if I were running a firm. And I mean, kind of the same thing with like audit PBC lists, client gets this giant file from you of everything you need and they don't submit it all. And sometimes they submit the wrong stuff, the wrong year, the wrong month, whatever it is. And you don't figure that out until it's like too late and then you're scrambling. Like you said, David, the AI agent could just do this every day, just go through all these clients and send the request again. Oh, we didn't get the right month. And, and AI can do that. It can compare what the client sent to what the list is that you requested and accurately figure out if it's correct or not. I have confidence that it could do that. So. David Leary: [00:50:57] Alright, so you solve all that. So the client intakes great. You're pumping out the return. Great. But if you still have a different bottleneck in your firm, you can't measure the throughput. Your throughput will not increase. Blake Oliver: [00:51:07] So that only works if that's the biggest bottleneck. So if the review is the biggest bottleneck, work on that. And maybe it's if review is the problem, it's having AI do the first round of review on the return and create like a detailed analysis of the return for the actual reviewer to help them. Yeah. Um, if it's I mean, that's all I can think of. That would be the two biggest. It's not the actual like return prep, putting the numbers into the forms or whatever. Uh, and in like financial statements, you know, I'm trying to think a lot of times that's also getting documents that you need actually, like you could do this for a cast practice or bookkeeping practices. If. You. If your bottleneck with a client is getting like the bank statements from them. Having like a scheduled task that runs at the beginning of every month and just. Ask your client for everything you need over and over again until you get it. And. That's something that we just skip as accountants because we are too busy and we don't follow up. And then client says, hey, it's the 20th of the month, why don't I have my books? And you're like, well, it's because you didn't send me all the stuff that I need to do it. And now they're. David Leary: [00:52:22] And you have a chance to send the email to ask. Blake Oliver: [00:52:24] And then they send everything on the 20th and their board meeting is on the 23rd. Right. So this has been a good, uh, this has been a good thinking session, right? It's like this, this is where the opportunity is for developers to build tools for accountants, because we know these bottlenecks and pain points. And here they are focusing on automating something that's like a pain point for them as an individual because it is time consuming to do your tax return as an individual. It takes a lot of time to put those numbers in the boxes, but we're really good at it. We're really fast at it. David Leary: [00:52:59] So why don't you read our next ad? Because that app maybe is took a bunch of processes and made it into 60s. Blake Oliver: [00:53:05] This is a new sponsor for us. New sponsor worthy with a you worthy says right now, delivering advisory services at your firm can take your team members 3 to 5 hours per client, pulling data, building spreadsheets, stitching together slide decks. That means your best people max out at ten, maybe 15 advisory clients before the whole thing breaks. Introducing worthy. Worthy takes you from completed books to branded client deliverables in just 60s, allowing you to turn every client into an advisory client without you building anything, hiring anyone, or changing how you work today. What used to take your team hours now takes any team member 30 minutes or less. Imagine giving your team that ten times the advisory capacity worthy connects directly to your client's QuickBooks or Xero files, and over 3000 other apps to generate insights across revenue, profitability, and cash flow using real numbers. Without hallucinations or these AI assistant Wes surfaces the insights to you first and never directly to your clients. The deliverables carry your firm's brand, not theirs. Your clients see a smarter, faster firm. You stay. The hero worthy is offering listeners of this podcast a full three months of access to the platform, including the ability to add unlimited clients. This is a great way for you to see how worthy can work for your firm and clients. To claim your three months of unlimited access, head over to The Accounting Podcast dot promo. That's The Accounting Podcast forward slash w u r t h y. Mail 22 on the YouTube live stream says that's why weekly reconciliations are important versus monthly. This is in regard to getting the client's books closed on time. I totally agree. And gosh, I'm thinking with client requests for like ongoing accounting work, you could like schedule this stuff out to go weekly with your tracker. David Leary: [00:55:03] That's what's, uh, who I forgot who he was. Like, that's his whole strategy. He build every week and he delivered financials every week. Blake Oliver: [00:55:10] Yeah, that was, uh, Jody Grunden Summit CPA's. Yeah. It's just like such a simple idea. It's like meet with your client every week. And they like that actually works without any AI. David Leary: [00:55:21] Manageable. The chunks stay manageable. Blake Oliver: [00:55:23] You have a scheduled meeting with your client every week. This is for virtual CFO services, but they're also closing the books, right. Scheduled meeting. You meet with that contact at the client and literally like they're on a call with you and you can go through that list. And if they don't get you what you need, it's like very clear, like that's the human way. David Leary: [00:55:41] And the list is like 2 or 3 things. It's not a list of 75 things at the end of the month. Blake Oliver: [00:55:45] And like, that really works because, uh, it creates accountability and stuff doesn't get delayed. And basically, I like that because you can kind of build the automated version of that for like bookkeeping, like send that request every week. Here's all the stuff we don't have that we need to close your books. I like this. All right. Well, we're almost at time. David, do you have. David Leary: [00:56:10] This one story? Blake Oliver: [00:56:11] Okay, go for it. Oh, yeah, you got it. David Leary: [00:56:13] So DHS employees are getting a one month tax extension. Don't know if you saw this or not. Blake Oliver: [00:56:17] Is this DHS Department of Homeland Security? David Leary: [00:56:21] Okay. They're getting automatic 30 day extension to file their federal tax returns. Moving their deadline to May 15th. And Treasury Secretary Scott Bennett framed this as a move, the way to ease pressure on the workers and families as they've continued to be affected by the government shutdown. And they continued their public safety duties during that time. Now I kind of see this and I'm like, is this just PR theater? Because they could have just all filed extensions and got to October 15th. Now, yes, apparently they'll get some penalty and interest relief as well. But like everybody can get an extension. I got an extension. They don't make a press release for that. But I do find some irony here. Think about this, Blake. So IRS budgets were cut. Those funds were used to they get shifted to DHS to make Ice hirings, right? Dhs gets funds get frozen because of Ice behaviors in Minneapolis or Minnesota. Right? And now the understaffed IRS needs to do some special accounting just for the DHS employees. And this was all preventable if you didn't cut the IRS budget and shift it into the DHS to begin with. And Ice. Yeah, like it's kind of ironic that like the understaffed IRS now has to do special work for the DHS employees who are using their allocated budget that they had before. Blake Oliver: [00:57:38] How about we just pay those TSA agents? Yeah. David Leary: [00:57:43] Well, apparently they got paid. I think they got paid. Blake Oliver: [00:57:44] Yeah, I know, but, like, why do we keep doing this to them? It's awful. Um, Dre says, Blake, do you offer mentoring services? Man, I would love to be able to, like, help firms with tech stuff. I mean, I love playing around with this. I just like I'm so busy with earmark and everything. Um. David Leary: [00:58:03] But retirement. Blake Oliver: [00:58:05] Retirement, it's like, and I want to do some courses on this on earmark. So if anyone's listening that teaches firms how to use coworker, these AI agents and wants to do like a webinar on earmark. We'd love to have it and we'd offer CPE for it. There just aren't a lot of people doing this because it's all so cutting edge. And I think it's, it's kind of hard to teach because like, yeah, this tool wasn't even on my radar until like a few months ago. Um, and you couldn't do this stuff just until like very recently. So it's going to take a little while to mature. And it's like, I totally get it. You know, I owned a firm. It's it's easy to have shiny object syndrome, but to actually build a tech stack that just works reliably year over year and, you know, generates a profit without you tearing your head out is hard. And you can't be using. David Leary: [00:58:57] Especially with things changing so fast right now. Yeah. You can't as soon as you build something, it's like out of date. Now you got to replace a different cog in a different cog. It's just we're in this crazy cycle right now. Blake Oliver: [00:59:06] And cowork, I mean, you know, it's a new app, right? It crashes. Right? It's like, I don't you might not want to have to worry about that. You know, on, uh, a few days before tax day and all that stuff. So. David Leary: [00:59:15] And I think that's what's interesting about a model like pilots pivoted again right now pilots, it's not a franchise model, but they are partnering with accountant local accountants that are calling it right in your town. But when I say it's like a franchise model, because basically you're going to get all of pilots, they're going to manage your tech stack for you and you just run your firm. Because I think that's it's interesting because right now, if you're a smaller firm, manage your tech stack and improving. It is crazy right now. It's just a firehose of stuff. It's really, really hard. Only bigger firms who have dedicated talent to do that. It's going to be the only way to do that. So maybe we're going to see a lot more, um, franchise type models or that, that you're, yeah, you're outsourcing your tech to a bigger accounting firm that it's. Isn't that the alliance model in a way for some of the alliances? Yeah. Blake Oliver: [01:00:05] It's like you get a, you get a. David Leary: [01:00:06] Tech stack. Blake Oliver: [01:00:07] From them and you get to have your own firm. Um, it's going to be really interesting. Like what an accounting firm looks like, especially a small one over the next ten, ten years, because we're going to be able to do so much more per person. But it's also going to require like a lot of technological expertise. So, you know, it's like, what, what's the, what are the roles in an accounting firm? I think, I think we'll end up seeing, well, this is this is is something, Peter, I was talking to Peter McCarroll earlier today and he said, I think we're going to have like a technologist role at every firm. And I could buy that. Right. Like you have the accounting tax expert, the CPA, and you have the technology expert, the technologist, and then the CPA is going to need like an apprentice, somebody to learn the business. So you could literally have like a CPA and, and, you know, a, an apprentice and you could have a technologist and then probably an admin and you could. David Leary: [01:01:15] So it's a little bit more like, obviously like the bench model where it was basically an accounting firm with engineers, but really it's probably going to be one accountant, one engineer and an admin person. Blake Oliver: [01:01:26] And like half a dozen. David Leary: [01:01:27] Person firm. Blake Oliver: [01:01:28] Just half a dozen AI agents that. David Leary: [01:01:31] Are. Blake Oliver: [01:01:31] 20 that are doing the staff functions. David Leary: [01:01:35] If you want to do that, I recommend following the Sastre blog because essentially that's their company. They have three three employees and 20 AI agents, and he talks about everything they're doing. It's really interesting. Blake Oliver: [01:01:45] Thanks everyone who joined us live on the stream today. Great to see you. Thanks for chatting with us. Um, love it. Love all your comments. If you are listening on the podcast feed and you haven't checked us out on YouTube, subscribe, go to The Accounting Podcast on YouTube and hit that subscribe button, that notification bell icon. You'll get notified when we go live. If you have a CPE requirement, you can earn free continuing professional education for listening to this episode and many other fine accounting and tax podcasts, go to earmark Dot app and your web browser, or download the free app for iOS or Android from the App Store. Free to sign up free to earn one CPE per week. And like I mentioned at the beginning of the show, if you've got a firm with five or more people who have a CPE requirement, you can get a 20% or more A discount off of earmark. You can get unlimited subscription for your team. And we also offer a private channel where we will create self-study CPE courses only for your team. From your recorded internal trainings and webinars. And David is muted. So I think that's. David Leary: [01:02:57] Sorry I had to open my door first. We're officially getting to summertime in Arizona where my studio is taking me. So I opened the door for a quick minute. I'm gonna have to turn on my internal AC. Blake Oliver: [01:03:06] We need to get you a mini split for your studio. David Leary: [01:03:09] Well, I have a I have a system, a fan system. I just didn't turn. I'm not used to turning it on. But now we've crossed a new line. I have to turn it on now. For now. Blake Oliver: [01:03:16] All right. We'll see you around here another day on on the live stream. On the podcast feed. Uh, David, have a great. David Leary: [01:03:26] Weekend this week, everybody. Good luck on your tax deadline by.