Solo CPA Firm Uses AI Agents to Handle 70% of Tax Prep Work

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] If people just use AI to avoid using an accountant, you're going to have clients you never even know you could have had because they're just bypassing you entirely. You're not part of the conversation. Coming to you weekly from the OnPay Recording Studio. Blake Oliver: [00:00:20] Hey everyone, and welcome back to another episode of the Accounting Podcast, your weekly roundup of news in the profession. I'm Blake Oliver. David Leary: [00:00:27] And I'm David Leary. Blake Oliver: [00:00:29] David, we've got two top stories this week. First, accounting today is out with their 2026 Best Firms for technology. And one firm in particular caught my attention. And that's because it's a firm of one one guy using AI to do all the prep. He just reviews. And he spends 70% of his budget on technology. We'll talk about that. And KPMG is shutting down their federal audit practice after the Department of War canned them. But first, David, let's thank our sponsors. David Leary: [00:01:04] Our sponsors this week. We have on pay. We have our cost seg, fishbowl and CNR consulting Group. So two new sponsors this week. Blake Oliver: [00:01:13] Let's thank on pay our lead sponsor. 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 coordinator who sets up worker profiles and transfers all the year to date data from previous providers. And that's all at no extra cost. They have a seamless QuickBooks and Xero integration that eliminates manual journal entries, and I can testify to how well it works. With zero, it's flawless. They also support any type of business. You serve farms, restaurants, nonprofits, you name it on pay can handle unique requirements without adding any complexity, and on pay keeps pricing simple to everything your clients expect, from multi-state filing to off cycle pay runs. It's included. There are no hidden fees and no surprises. To book a demo, head over to The Accounting Podcast dot com slash pay. The Accounting Podcast dot io forward slash OPAY. And now, David, let's talk about that firm of one Accounting Today did their annual Best Firms for technology list, and on this list is a firm from Richmond, Virginia, called the millennial CPA one office, zero employees, one partner and a budget for tech of 70%, Which is, uh, really high. That's a lot. 70% of your budget is technology. I mean, it makes sense. You don't have an office or you work at home, right? David Leary: [00:03:05] You don't have you don't have a staff, no employees. Blake Oliver: [00:03:07] Why not spend it on technology? And what kind of technology? Artificial intelligence, of course. I'll just read this profile. Artificial intelligence allows firms to do more with less, but the millennial CPA takes it to a whole new level. It's a sole proprietorship. One man, Sam Leon, runs it all. And AI has been key. The millennial CPA is a tax firm first and foremost, and AI does most of the work while he, the human, does the review to make sure the model got it right. While this is a relatively new kind of tax workflow, he believes it will become more common as time goes on. I see AI as coming together to be a total tax preparer, and whoever signs the returns is the reviewer, he said, adding that it can even create its own workpapers once the human feeds it. Structured inputs. It has been especially helpful with tax return review by including redacted tax information and asking for year over year comparisons. Leon said work that takes a human hours can be completed in a few minutes. Additionally, creating detailed tax workpapers from QuickBooks exported financial statements in a highly interactive Excel export, which would have taken a human 3 to 5 hours, takes AI five minutes. Of course, Leon always has the final say, and everything the AI produces is reviewed by him. The bot is better thought of as a junior preparer, he suggested, while he would be more like the senior team member. As for now, this is how he expects things to remain. Essentially, the strategy is I won't hire until I hit a wall with my AI preparers and AI workflow managers and what they do for me, he said. David Leary: [00:04:39] So does it go any detail? So obviously the work, the actual doing the return, reviewing a return, signing off on a return, he's using bots and AI agents to do that work. That's very clear in this article, but there's a lot more to running a firm than just that. Like, how is he interacting with customers? How is he getting document retrieval? How is he getting people to pay and invoice them? Is that all agents as well? Does it does it talk about that in this firm? Blake Oliver: [00:05:02] That was the entire profile. So you really want to. David Leary: [00:05:04] Have to get this on the show. You got to invite him on. Blake Oliver: [00:05:06] I want to have him on the show. I mean, I think this is all totally doable right now. Like I'm using cloud Cowork every day. You could do all the tax return prep with that tool, drop all the client docs into a folder, or just point cloud cowork at the folder in your file storage, and it will organize all the source documents, rename them, put them into folders. It can then create the like a document with all of the key information, like a spreadsheet, right? With all the information you need to put into the boxes, into the tax return software, you could use it like he said, to take QuickBooks financial statements and then do the trial balance mapping that you have to do in order to get the financials into the right mapping to put into the tax software. David Leary: [00:05:56] Which which I think I've seen half a dozen apps that have popped up over the last eight years that are trying to solve that use case. And I don't even know if they ever finished and got to market or what. Yeah. Blake Oliver: [00:06:07] Something else you could do with it right now is point that cloud project at the client folder and give it the prior year return and ask it to figure out what's missing. What did the client not provide? Did they provide the correct year, the correct document? Is it legible? All that kind of admin work. I mean, you could create the client, uh, list of documents you need from the prior year return. Also, just put that, you know, get that return in there. Um, I mean, so many possibilities. It's hard to do at scale in a firm because right now, Cowork is a tool that everybody configures individually. For the most part, you can share skills among users in your organization, but to actually like do it at scale is hard. But to do it as an individual is totally possible. And so I expect we'll see more of these firms of one, and you'll be able to scale up and make a lot of money, because you don't have to hire employees. I mean, that's your biggest cost in a firm and be really profitable. Tait, in the live stream, asks how much revenue does he make? We do not know. I want to find out from Sam. Let's see if we can get him on the show. And Tait also asks, is he a software engineer? We also don't know that, but I don't think you need to be one in order to use these tools. Cowork is pretty simple, relatively speaking. I mean, you don't you don't need to know any code to use it. You just need to know how to prompt AI. And it's getting easier as the models get better and better and better. David Leary: [00:07:34] Take your existing skill set and let you exponentially increase it. And I had an article from. It's been sitting in my stack for a couple of weeks ago, but the concept of a minimum viable sized company. And it really is. It talks about what this person's doing, right? You can have these teeny teams build a company and you can have growth because the old model, if you wanted to grow your company, what did everybody do? You go raise money and hire a bunch of people and that's how you grow your company. You just get lots of humans to build your company bigger and you don't need that anymore. So the future, the winners are going to be like you just described, small, highly efficient teams with a strong strategic clarity. Not large orgs, right. It's going to be small individuals using AI to level up to 20 X their productivity. Blake Oliver: [00:08:21] Hk geek in the live stream on YouTube says between anchor and double, you can automate so much of the communication proposals and accounts receivable for your firm. I agree, and by the way, you can now connect double to klod via the double m CP server. And I've done this. And you can use Klod to help you with your month end checklist, your closed checklist and double it can go in and pull it and then help you work through the reconciliations. And it's not perfect. It's a little clunky still, but I've even got it opening up a tab in Chrome in zero and doing the reconciliations. So if I dump the bank statements into a folder project folder, it can go pull all the beginning and ending balances and see all the transactions and import them as a a bank feed via CSV import into zero if it needs to, and then run the reconciliation report and check everything off and look for errors and troubleshoot if there's a reconciliation discrepancy like that is powerful. Now it's slow because clicking around in a web interface is something that's very slow and uses a ton of tokens, but I think