Earmark Podcast | Earn Free Accounting CPE

Attention: This is a machine-generated transcript. As such, there may be spelling, grammar, and accuracy errors throughout. Thank you for your understanding!

Blake Oliver, CPA: [00:00:00] My guest today thinks some accountants are about to outearn software engineers, and she's a software engineer herself at her bookkeeping, tax prep and CFO advisory firm. The target is $1 million in client fees per accountant. I'm Blake Oliver and this is the Earmark podcast. Jessica McKellar studied computer science at MIT, co-founded pilot as its CTO, and now runs it as CEO. Pilot sponsored this episode. Here's the question at the heart of it. If software runs, the clothes and accountants pick up it, review that efficiency is worth real money. So who gets it? The firm, the client, or you? Jessica thinks some of it comes to you. I think it might all come to you if you're a firm of one. Here's Jessica McKellar. So Jessica Pilot closed nine years of books. That's something like 187,000 Months. Month end closes before you started selling this software. Meridian to anyone else. Why did you spend nine years running an accounting firm before you started this?

Jessica McKellar: [00:01:15] Yeah, this was definitely not the lowest pain way to validate that we had solved this problem in the way that we wanted to. We did it because we thought it was. It was the only way to make sure that we had built the right thing, and that it would be effective and delightful to use at scale. And so, yeah, if we had this original hypothesis ten years ago of it is possible to substantively automate bookkeeping and to use that as a foundational automation layer to run a firm on. Yeah, you know what, it took us almost a decade for a for us to build out the tech and to prove it to ourselves. And just candidly, for some of the kind of frontier AI stuff to mature to the point where the interface that we're able to provide to firms is so effective.

Blake Oliver, CPA: [00:02:06] So Meridian is the tool that you were using internally at pilot, and that you still use at pilot to close the books for your clients. So tell me about it. What is it?

Jessica McKellar: [00:02:21] Yeah, totally. So and so pilot is like the test lab for Meridian. Meridian is an AI accounting platform for accounting firms. And there are a couple of key pieces to Meridian. The core of it is, number one, the ability to articulate recurring work in a structured way, like in a checklist, and to have the system autonomously execute that work. So, for example, a month end close process where the close is done beginning to end, like down through to review, uh, like in software that is Meridian that works in concert with the ability to direct Meridian to execute ad hoc tasks as well. So you can think of like the core automation of Meridian is you can articulate recurring work in a checklist and have it executed autonomously by, you know, it's going to feel like a really competent assistant to you. And then you can also direct that competent assistant to do to, you know, action, review feedback, respond to client requests, do complex accounting tasks like revenue restatements or what have you, um, you can you have this chat interface to that as well. That is like the core of the automation. There's also a client facing portal for firms who want to use it. And then there's like the firm management piece of this. And that's like maybe the, the four pieces together that at the highest level describe Meridian.

Blake Oliver, CPA: [00:03:51] Okay. So when I'm in there and I'm looking at it, I'm seeing a month end close checklist. So I'm picturing bank recs, uh, payroll, journal entries, accruals, deferrals. Is that what we're looking at? Is that what we're talking about?

Jessica McKellar: [00:04:07] Yeah, totally. So, you know, this is you're closing the books for arbitrary business, arbitrary business model, arbitrary industry. And it's like from the very beginning to the very end, we need to first have all the data inputs to have an accurate month end close at a minimum. That's your bank and credit card activity, but usually it's other stuff as well. Maybe there's a payroll sync, maybe you're rolling forward a loan schedule, you're booking revenue via some source data, etc. so it's, it's like step one is get the input data. Step two is execute the close checklist. This is a checklist that could be the out of the box Meridian checklist, which is like quite a, it's like an effective accrual basis close for many firms. Or you can have a firm level closed template, or you can have a client specific checklists like customizable down to the task task level. But yeah, there's a, there's a checklist that is executed starting with transaction import and reconciliation through all the things that you would expect, categorization and dealing with AP and AR and doing the rev rec and booking payroll, booking all the journal entries. Oh, Meridian also is maintained. So all of the kind of supporting material that you would want maintained as part as part of completing the month end close is also built natively into Meridian. So Meridian is maintaining it can inherit, maintain and roll forward supporting schedules. It's maintaining the recons. It's maintaining the tick and tie groups for clearing accounts. It's got a vendor mapping and categorization history. So it's maintaining all of that as part of completing the close. And then it's posting directly to the GL. And so the end result of a Meridian close is you have completed books, you have updated schedules, and you have like client work papers. And then the accountant responsible for that customer's clothes has a really tidy review summary to take as an input to their holistic review.

Blake Oliver, CPA: [00:06:18] All right. So and the work of like all of these tasks is, is no longer being done by humans, but it's being done by AI agents that take on each task.

Jessica McKellar: [00:06:33] I mean, yes, definitely. I mean, there's a lot there's like a lot of deterministic software that is like sort of being applied under the hood. There's a bunch of sort of aggregate intelligence from having exercised these capabilities across, you know, many thousands of businesses across many tens of thousands of closes. But yes, there's a bunch of tools, software and sort of aggregate context and intelligence. There's a bunch of firm level customization and preferences, client level preferences. And then there is a kind of a purpose built AI accounting agent framework that is executing this closed process.

Blake Oliver, CPA: [00:07:14] And how automated is this? Like, like I'm, I'm comparing this to, say the current process, which, you know, if I'm a well organized, oiled machine of an accounting firm, I've got my closed checklist for every client. I've got my staff working through all the items and getting those all taken care of. Like you said, updating the supporting schedules, getting the documentation from the clients, etcetera, etcetera. Um, like what, what does Meridian do in terms of the, the automation, like what's it capable of?

Jessica McKellar: [00:07:47] Yeah. I mean, so what we've built and the experience that we, that, we want to provide. And it's important to I think it's, it's an important choice that we've made that this is the abstraction that we provide. It's as long as the input data has been provided, the sinks are live. The files have been uploaded. The close is executed autonomously beginning to end, like all the way through to being ready for review. So, um, there are situations where it can, it can truly be blocked on like a data access issue, for example. Um, but generally speaking, if you have like the well-articulated data inputs and a well-articulated checklist, it will, um, it will work. It's going to work through to the end. You can, this is configurable by the way, I mean, some, some firms want really sort of granular review of major steps in the process. You can configure it to be as babysat as you want it to be. I think most firms find that once they get comfortable with it, they're pretty stoked actually, that you can just get sort of kicks off at the beginning of, of the new month and um, you wake up to completed books.

Blake Oliver, CPA: [00:08:59] I mean, I have to, I have to say, I've always been skeptical of the claims of like fully autonomous month end close because you mentioned it. One of the blockers that comes up in almost every task is, oh, I have a question, or I don't have this piece of information that I need from the client, and then I have to stop my work and go ask for it and then wait and follow up until I get that and then come back. And then I've forgotten where I was and then continue. And that's often a big reason for the delay, both in month and close, but also in tax in every area of accounting. Practically, we just never seem to have all the information. So how does Meridian handle that? If it if it's working through a close task and it has a question, what does it do?

Jessica McKellar: [00:09:43] Yeah. So there are some kinds of situations that are truly blockers. And then there are some where some judgment can be applied. And then that can that kind of that judgment call can be surfaced during review for the, the person responsible for the account to double check. And there are a lot of cases where it's like, it's not, it really kind of can and should be non-blocking. You see some kind of new behavior from the business. They, they, you know, a new kind of.

Blake Oliver, CPA: [00:10:22] A new transaction comes in.

Jessica McKellar: [00:10:23] Activity. Yeah. Like we've got.

Blake Oliver, CPA: [00:10:25] We've got a new expense.

Jessica McKellar: [00:10:27] Yeah. They appear to have received some financing and there's like no prior history on that or yeah, you're seeing payments to, you know, you see a payment to a new payroll system provider and, you know, like don't have the payroll access, right? These are situations where you can make a call that's reasonable, that produces reasonable draft books, and then there may be a process to, you know, you may need to go ask the client for information. You. It's going to be like a call for the person responsible for the account of like, do we want this to literally block the close? We're going to like go make the client request. We're going to wait on their feedback or are we going to ship, you know, partial financials to them. Uh, and then once they get us the information that we need, then we'll like book that and then and move on.

Blake Oliver, CPA: [00:11:16] I get that.

Jessica McKellar: [00:11:17] Situation dependent, but, but the, I think the thing that matters the most is when do you truly have a blocking issue here? Like I don't have access to the to the primary checking account. I don't really have a way to materially complete the books versus, oh, we can like park this somewhere reasonable. I can FYI this as a call that I'm like, I, the agent made as part of the review summary. And then we'll let the reviewer decide if they want this to block the shipping the books or not.

Blake Oliver, CPA: [00:11:42] And then when I'm reviewing it, I can chat with the agent and then give it the information it needs and it can go and adjust things.

Jessica McKellar: [00:11:50] Yeah, totally. So just to take a trivial example here, let's say you're working on a client and they, um, you see payments to gusto for the first time. So it's like a payroll provider that is being paid. You haven't seen this before. Maybe they're like, they've switched off of ADP, I don't know. And we don't have access to gusto. So probably what's going to happen as part of the close is these are accrual basis books. Generally speaking, you want to be like booking the cash out to the like a payroll liability account or set of accounts. And then you want to be booking the granular like payroll liability breakout, the wages and the taxes and whatever, like out to the PNL. Great. If you don't have access to that data, if Meridian doesn't have access to that data, probably what it's going to do is it's going to park the activity either on the, in the, in the liability account or like on the PNL somewhere reasonable and call out to the reviewer, hey, until we have access to the payroll system or someone uploads a payroll journal, we don't have the information to do the granular breakout, so I've parked it somewhere reasonable. Let me know if you want me to park it somewhere else.

Blake Oliver, CPA: [00:13:00] Just like, yeah, I would park that in like a payroll expense account with a note that, hey, we need to break this out into different expense accounts once we get the information right. Bonuses. Exactly. Wages, all that benefits. Interesting. Um, so, so like, what is, what is the impact of this at pilot? Like you're using this, you, you started out pilot, uh, like, what was it eight, nine years ago? Yes. And you're doing it all manually. I mean, to be sure when you started, right? Oh. So yeah, yeah. And now, uh, thanks to the, the, this AI renaissance we're having, you can build the, the automation in. So I'm curious to know what is, what is the impact of that on your labor?

Jessica McKellar: [00:13:49] Yeah, totally. Well, if I in case it's interesting, let me slightly over answer this question because you're totally right that the way this started was there's zero software. My co-founder, who I've known forever, and we met as computer nerds at MIT in the computer club. And we've worked together ever since for almost 20 years. He's sitting at QuickBooks online. We've, like, begged our friends to let us do the bookkeeping for their businesses to be like guinea pigs. He's like manually clicking around in QuickBooks online. I'm like hovering over his shoulder with my laptop, like, right. Like manually. This is like back in the stone ages of 2016. I'm like manually typing in a bunch of code. Like I'm like implementing the first version of the software against the QuickBooks online APIs, right? So it starts with like nothing except manual bookkeeping. And we spend a bunch of time building out the sort of initial, the first generation of this software. That was cool. We described it as like the Iron Man suit for accountants. It was like a bunch of tools that an accountant or our accounting team could kind of plug themselves into that would execute pieces of the clothes and effective, but ultimately an experience that looked like a bunch of context switching or handoffs between a person and the software on getting the work done. Right. And there's like a bunch of back and forth and a bunch of babysitting, and it's like, nice to be able to have, um, tedious tasks automated.

Jessica McKellar: [00:15:27] Uh, but it can be challenging to turn that experience of like going back and forth with the tools into a day that is delightful for an accountant and like a day that is materially more productive if you're, if you're like a firm owner who's like trying to move firm level metrics, it can be surprisingly difficult to achieve that if it's like workflow augmentation, right? And so the thing that has happened in the last, I would say 18 months is the foundational models have advanced to a point where you can, you can actually take the accountant out of the Iron Man suit and you can put, uh, an like, like AI or an LLM into it. And then it can actually like drive with all the tools available to it, with all the contacts, with all of this deterministic supporting schedule, like vendor mapping support the tick and tie support the recon support with all of those tools and capabilities. Now you can have an agent like do all of the last mile thing. And so that's what gets you the beginning to end autonomous clothes. And that was really the inflection point where we said, okay, we've got an interface now that is actually simple, delightful, something that accountants that are not weirdos at pilot HQ would actually want to use. Um, and that this can actually be transformational for the firm. And then like the things that actually, like, in some ways matter the most about this are it's like, okay, anybody who's, I mean, anybody who's worked at an accounting firm, if you've been through like, you know, years of month end closes or you've been through years of tax seasons, there is so much of like, oh my God, is it? Are we having the mad scramble again? Is it crunch time again? Do we have a capacity shortfall again? Am I working on Sunday again because someone was out sick and we're all rallying to pick up their work? The thing that this really does, that if you have an autonomous month end close, that's predictable in software.

Jessica McKellar: [00:17:41] And the same would be true for like a tax process or a PR or processing or there's like other recurring work that we could talk about. But let's focus on the clothes for a second. If you have like a predictable autonomous month end, close it like reduces the work, but it reduces the variance and it makes the there's so much of the stress melts away if you don't have all of the capacity mismatches, that can come up with the variance of trying to do all of the recurring work that needs to be done when, you know, unexpected stuff happens with the team, and then unexpected stuff also happens with your clients. So that is like the super magical thing that I at least have experienced with Meridian. You get this predictability that you can like plan your time around as a team.

Blake Oliver, CPA: [00:18:31] You get predictability because it's doing these tasks the same way every time. I mean, you can customize them, right? But if you want, it sounds like you can have like a reconciliation done the same way consistently across every client. Is that right?

Jessica McKellar: [00:18:48] Yeah. The the variance goes down in part because it takes less time. Like just the, you know, this AI assistant with infinite time and patience is like, like work that is, you know, uh, time consuming, but mechanical is like delegated. Um, the, you know, the, basically all of the month end close is delegated up through to review. That review process tends to be like lower variance than like, how long did it take to actually like do the production close work? And so it's like the total time that people need to spend on the close is much, much lower. And also the variance in that time is much, much lower.

Blake Oliver, CPA: [00:19:27] One of the concerns of autonomous AI and, and turning these tasks over in the past and in the really recent past has been accuracy. Totally. Because AI has. It's a these, you know, gen AI is statistical, it's probabilistic. And it it's guessing at the right answer. We've seen its accuracy improve. But in accounting, you know, 90 something percent is often not enough. It needs to be 100%. We need to tie everything out. So how do you ensure enough accuracy to let this system run on its own through a close without the human oversight like during it, the human babysitting it?

Jessica McKellar: [00:20:13] Totally. And I, you know, I think this is this is a place where I think Meridian has been architected in a very intentional way that is perhaps distinct in the market right now. So first, there's a lot of determinism in Meridian, and there are there is this native understanding, a bunch of the core concepts that you would need to have the system prove to itself and prove to you, and make it easy to prove to the client that the work was correct. So this is like reconciliations, right? Maintained in Meridian. This is the supporting schedules that are maintained in Meridian. This is like the tick and tie visibility for accounts that have that tick and tie or clearing account property. This is like the vendor mappings. This is all this is not stuff that like that's a really different experience from like plugging a client checklist into Claude and seeing what happens. There's like a lot of longitudinal data history that is being applied in completing the current close. Um, there's a bunch of deterministic checks that are executing. Like if you take reconciliation as an, as an example, like as part of this, you're the system needs to be able to reconcile against the source of truth for your bank and credit card and other like balance reconcile accounts. A you have that, you can see it, you can see what ladders to that balance. And then if it if a, if a recon breaks, it knows about it and will flag that.

Jessica McKellar: [00:21:53] And it's, you know, if it's self, if it's something that can be self-healed, then great, but often that represents something that actually needs to be addressed by a person, often in correspondence with the client. Because, you know, maybe what happened is the client posted a bunch of invoice and bill payments to the bank feed, and that's what broke the recon. Like what was their intent behind that? So you have like the guardrails of the determinism of these core accounting subsystems. The same thing is true with supporting schedules. Like it's not like you're spreading out a vendor payment over a service period in isolation. This is like being registered in supporting schedules that are then being deterministically rolled forward. And you can look at them in the UI and you could change them from the UI if you wanted to, or you could direct the agent to do it via the chat interface. But you just, you have, you have like all of that proof of work that you would want to see. And then, and so, and then it's very like self, like all of these steps, like major steps in a close process are checked internally by the system and then ultimately surfaced for a like holistic review by a person that is like the ultimate check and balance on this.

Blake Oliver, CPA: [00:23:03] So let's, let's, uh, think of, I'm just thinking this through. So there's like a supporting schedule for some sort of, uh, you know, uh, account balance, right? And the, the, the end value on the supporting schedule has to match what's in the GL. So there's like a check that is happening. So if the AI builds it or rolls it forward and it like messes it up, then it's checking to make sure that the end balance ties out, and if it doesn't, then it just goes and it goes and fixes it itself. Right. Is that what you're saying?

Jessica McKellar: [00:23:41] It's actually an even stronger guarantee than that in that it the AI is not like being told, go roll forward the supporting schedules. There's like deterministic software that is doing the math of rolling forward the supporting schedules. Okay. Um, yeah, but then yes, but, but, but then the, but then the AI agent is gonna verify that everything matches. And if something doesn't match because you know why? Because for some reason, a client deleted a journal entry, right? It'll, it'll see that and surface that it's potentially clear what the remediation is. Or maybe that actually needs to be a conversation with the client.

Blake Oliver, CPA: [00:24:21] So one of the issues with using like Claude Cowork, for instance, to do the month end close is that I can do it right. I've figured out how to build skills that can then go into my GL and can, you know, complete bank recs and stuff like that. Um, but like one of the issues with scaling that up in a firm is the, the audit trail because, you know, I've got to be able to understand why the AI did what it did, why it categorized the transaction, um, all the reasoning the same way. If you know, like if an auditor came and asked me what I did, I'd know it, right. And hopefully I'd have documented it and I'd be able to explain like, how does, how does Meridian handle all of that? Like, because if I'm using it right, I want to know that if I get asked to, to show proof of work that I can.

Jessica McKellar: [00:25:12] Yeah, yeah. And probably what you would expect here, there's a full audit log. And like it was just visible in a couple of ways, like within Meridian also, it's also posting out to the GL. So there's like the GL audit log. There's what you have in Meridian. And it's not just like what is the audit log on individual transactions. It's like this transaction was categorized here. Why? What was the categorization? History. This is like what's going on with reconciliations? What's going on with ticket groups? What's going on with the supporting schedules? You have all of those artifacts. You have the monthly cuts of those artifacts. So you have the history of the closes as well. So there are a bunch of a bunch of altitudes at which you have a bunch of sort of evidence or support for why something is the way that it is. And it's like all in a nice UI too. It's not like you're having to do a bunch of forensics within with an LLM to figure this out. You just like look at it in the UI if you want.

Blake Oliver, CPA: [00:26:15] We've been using the term general ledger generally, but we haven't actually talked about which ones. So I mean I Meridian is not a GL right. It's what is it. What is it working in.

Jessica McKellar: [00:26:31] Right. So I would describe Meridian in one sentence as an operating system for accounting firms. And you. Yeah, it's gonna need a backing GL for each client, right. So the client has whatever they're using QuickBooks or they're using zero or they're using whatever ledger they're using. And then Meridian is like hooked up to that ledger and is both pulling data from the ledger and is posting directly to it.

Blake Oliver, CPA: [00:26:57] How do you deal with like limitations on what those companies will actually let you access in the GL? Because one of the issues I've run into with using Clod is that I can't get to everything through the API. So even if I or the MCP. So even if I connect clod to zero, for instance, it actually can't do very much. It has to open up like a browser window like me and click around, which is really slow and inefficient. So how are you handling those limitations?

Jessica McKellar: [00:27:24] Yes. Um, and I, let me stop me if I go too far in answering this, because this is an area that I think is really interesting. Um, so I mean, one way of thinking about Meridian is that it actually sort of provides a universal abstraction over the clothes that is ledger agnostic. So if you think about QuickBooks online, great. Actually really open API comparatively, like the good and the bad of that. But like none of the supporting schedule support within QuickBooks online that some of the other systems or the more up market ERPs have. Right. So if you have clients that are like in QuickBooks online, you're probably maintaining supporting schedules yourself somehow, right? Mhm. Um, different from zero zero has a very strong opinion about the bank feeds and reconciliation and does not expose that via the API. You have to do it via their UI and they have like partial supporting schedule support because they have like a fixed asset subledger, but nothing else. And then like, next was a totally different story. And Sage Intacct is a totally different story. So every ledger is a little bit different here.

Jessica McKellar: [00:28:35] And yeah, I mean, honestly, we've just done a bunch of work to understand this and abstract over it. And it's true that depending on what ledgers are common in your firm's client base, there are some things that. The only option is to do some stuff in a browser, like the zero like bank feed recon matching process is an example of that. Um, what's kind of, I mean, I think this is going to continue evolving because I think that these ledger providers are going to need to be responsive to the market wanting to be able to automate against their APIs. But like today, for example, if I were closing some Xero books in Meridian, I would have a workflow set up where I'm actually driving Meridian out of, uh, like Claude coworker Codex with browser automation enabled. And so I can actually still have that experience of the autonomous close, but it is being driven off of my computer. So I get the browser use with my like account authentication for that part that has to be done via the UI.

Blake Oliver, CPA: [00:29:45] So I can connect Claude to Meridian.

Jessica McKellar: [00:29:48] That is something that I can do and is not something that you can do. But why don't we check in in weeks, weeks to months? And then we're there's, we're shipping a lot pretty quickly on this.

Blake Oliver, CPA: [00:30:00] Yeah, that's really interesting because like that has changed. Uh, my, the way I work all these connectors that are now available, right? I've connected all of my systems and they're all adding MCP servers, which for our listeners is basically the way that these AI agents talk to each other. And so now, you know, I can do some like really impressive stuff. Uh, not even I can just chat with my clod or ChatGPT instance, and then it goes and does the work for me.

Jessica McKellar: [00:30:32] So yeah, the magical trifecta is for me is like, I tend to use Codex, not cloud, but it's like Codex from, from OpenAI. It's like Codex Meridian and then my email and then we're, we're rocking and rolling when these, when these three systems can talk to each other because I can just be like, hey, this client, you know, sent me this like really long, complicated set of questions and like requests for restatements in this email Meridian go, please go take care of this for me. And I just do that.

Blake Oliver, CPA: [00:31:04] Yeah, exactly. It's linking up the communication that has to go back and forth with the, the system for getting that work done. And I feel like that is such a huge opportunity because accountants, we're over here on one screen in our email, and then we're over here on another screen working. And these things are so disconnected. And we have to be processing all this information coming in, you know. And that's the overwhelming part is totally. And then we can't be responsive. And that's what clients complain about the most is, hey, my CPA doesn't get back to me on time. It's like weeks later.

Jessica McKellar: [00:31:40] So totally. And I think and Meridian really, um, supports, we really want to meet firms where they are on this and it really supports two major models of work. The first is it's, it's like the only thing you have to worry about is Meridian. It's like within the walls of Meridian have access to everything that you need to have access to. And like batteries included, you can do as much configuration as you want. But actually the out of the box experience is like very effective. And then for folks who have their own way of working that they like, that they want to compose Meridian into supporting that as well. So if you want to like come into the Meridian garden and like play there, that's great. Or if you have your own workflows that you want to bring Meridian into, we want to support that too. And so you'll see that being possible via. Yeah, like the way that you can compose work with plugins for the, the LMS as well as like what's batteries included inside of Meridian.

Blake Oliver, CPA: [00:32:39] Oh that's fantastic. Um, so, uh, my, my podcast co-host, David Leary, we were just talking about that this, this morning on, on our live stream of the accounting podcast. We were talking about into its earnings call. And they are talking a lot about adding AI into QuickBooks. And so, you know, uh, Meridian layers on top of the GL. And I'm going to assume that most of the time it's going to be QuickBooks in the United States because that's the 90% market share.

Jessica McKellar: [00:33:12] Yeah.

Blake Oliver, CPA: [00:33:12] Yeah. So with, with Intuit building their own AI agents into QuickBooks, uh, you know, why don't, Do you think they're just going to build this? Like, is that a threat to you? I mean, like what? What's what's to stop that?

Jessica McKellar: [00:33:33] Yeah. So I think it's a really interesting time to be a ledger. I think it's an interesting time to be. Yeah. One of I mean, I don't know what Intuit would want to call itself. It like, you know, kind of seeks to be the platform within which all work is done. It's your CRM and it's your payroll and it's your accounting and it's your everything. Um, really interesting time. Um. What I care about and what we're trying to build with Meridian is that actually the heart of this goes back to the end, the end business owner in particular, like the small businesses of America and me wanting them to have access to high quality finance and accounting support, which they're often priced out of today. And so can we equip firms with an accounting platform that. Opens up access to opportunity for the end business owner. That's why I'm coming at it from that perspective. And it's a big market. And Intuit is busy doing what it's doing. And Meridian is busy doing what it's doing. I don't I'm not. I worry about providing a great experience to firms and to their end clients. Um, I do think that. I think it's going to be important for these ecosystem players to like, decide if they want to be encouraging folks to build on top of them or if they're not. Like they're either going to want that or not want it. And I think that people are going to migrate to the systems and tools that work for them, for the way that their workflows are evolving. And so I just, I think that companies like Intuit are certainly paying attention to that and will have to be mindful of and careful about that, because people want to be able to automate customized, build their own workflows on top of these systems. And it's like Meridian is an example of that. But like also, you know, companies are automating their workflows internally, right? Um, like lots of people are building interesting workflows on top of these ledgers and ERPs.

Blake Oliver, CPA: [00:35:46] You and I are the kind of people who tend to just love automation because we don't like doing boring work. I can tell, and I'm sure many of our listeners feel the same way. Um, but also many, you know, buyers of Meridian, many accountants are doing things a certain way without that automation. It's been working great. You know, they are not really interested in changing things unless there's a really good reason. And one of those reasons is often the return on investment. You know, if I can make more money, if I can cut costs, increase profit. Like we're accountants, we think in terms of that PNL. So setting aside just the lovely side effect of not having to do this boring work anymore, what's the what's the ROI case? What's the return on investment of putting something like this into a firm?

Jessica McKellar: [00:36:41] Totally. And first, I'll say that I'm like not trying to. Different firms have different goals and priorities, so I. It's. For some firms, investing in workflow changes is a priority right now. And for some firms it's not. And that's totally it's great. It's good to know yourself and know your priorities. Like, why are a lot of firms interested in evaluating AI tools and platforms right now? Uh, well, I think it's because right here from firm owners that it's because, uh, they want to ensure that they're delivering a great client experience in a way that is priced competitively in a way that, um, supports strong firm economics, but also a really compelling environment for their accounting team or finance and accounting teams, like in terms of the comp and benefits as well as the day to day experience. And I think a lot of friends are finding that some innovation is proving to be necessary to make the math work out on all of this. It's like end client expectations are rising. They also all have access to Claude and ChatGPT and are seeing what they can do themselves. Like there's rising client expectations. Uh, you know, I think labor costs continue to, you know, shift over time. The firm wants to make sure that it is set up for success and able to scale if it wants to. Some firms also want to focus on, um, yeah, like reducing the stress, reducing the crunch time.

Jessica McKellar: [00:38:30] Can we have a four day or a three day workweek instead of a five day workweek? Some firms want to scale and some firms want to get, you know, take it easy. Um, so I think different firms have different goals. I think a lot of firms are finding that the basic math is starting to break down. And so they're looking at tools that can help improve this equation for themselves. And like, so how does Meridian plug into that? Yeah. Well, if you have the production closed done, like completely end to end in software, all the team needs to do is a review that is a lot. That is typically a lot of margin dollars that can be reclaimed and reinvested in the client experience in your employees, or just like accruing back to the firm economics. And again, different firms leverage this in different ways. And then on top of that, I really can't overstate how great it is to. Not be stressed out as an accountant on the last day of a close where you're like, guess what? All my client work is done. There were no curveballs, there were no capacity issues. And like, you know, Meridian meaningfully impacted that experience because so much of the production work was done on your behalf because, uh, the tedious work was all delegated because the variance went down and you were able to plan your week around that.

Blake Oliver, CPA: [00:39:53] So well, give me, give me the numbers.

Jessica McKellar: [00:39:55] It's never been more fun to be an accountant, I tell you.

Blake Oliver, CPA: [00:39:58] Give me the numbers. Right. If if if Meridian is reducing the amount of work that I have to do or my staff have to do, closing the books, that's going to reduce my, uh, my cost of labor, right. On a, on a per client basis. Or maybe it's going to let me take on more clients and make more money. So like what is it? What is it actually doing? Because you're doing this a pilot, right? You're seeing the results. Like what is the impact?

Jessica McKellar: [00:40:23] Totally. Yeah. And again, every firm is different, has different priorities at like a pilot as test lab for Meridian, uh, we goal ourselves internally on $1 million client revenue portfolio per accountant.

Blake Oliver, CPA: [00:40:39] Oh, I just got to stop you there. Okay. A million in revenue per accountant.

Jessica McKellar: [00:40:45] Yes.

Blake Oliver, CPA: [00:40:46] That's a lot.

Jessica McKellar: [00:40:47] It is a lot. And you know what? Just think about how great that could be from a compensation perspective for accountants. If some of that efficiency is getting accrued back from a compensation perspective, that's really I cannot wait to look. I've been on both sides of this. I am a software engineer by training, and I have now spent ten years doing like a lot of bookkeeping in particular, I think that we're about to enter a period where accountants are going to start making more than software engineers in some cases. And that's pretty interesting. And just like zooming all the way out from to like the big, big picture, like there's like a lot of, there's been a lot of like freaking out and doom and gloom about like accountants shortages. This is going to change. I think as, as like the comp profile of this discipline really changes. And I'm like really excited for that because there's a lot of really talented, passionate, capable people who are about to be able to delegate a lot of this mundane work and focus on even more leveraged activities. And it's going to be really fun.

Blake Oliver, CPA: [00:41:59] Yeah, we talk a lot on the show, on the accounting podcast about, um, this idea of the firm of one being a big deal in the future because, you know, if I'm one person running a firm solo CPA, solo tax preparer, solo accountant, whatever. Bookkeeper. I mean, if I can generate a million in revenue and not have to hire anyone, that is, uh. Well, I mean, like you said, I'd be making more money than all the engineers except the ones working for, like, OpenAI and anthropic. Right. And and meta. Sure. I imagine, you know, like I've just heard those crazy numbers. Yeah. So let's talk, let's dig into that. Uh, sure. How AI has changed engineering and computer programing and software engineering. Like you are a technical founder. You were the CTO at pilot, right?

Jessica McKellar: [00:42:57] That's right. I started as a CTO and I moved into the CEO role. Yeah.

Blake Oliver, CPA: [00:43:00] So see.

Jessica McKellar: [00:43:01] Sam founding team, we did a little bit of a shuffle, but yeah, yeah.

Blake Oliver, CPA: [00:43:04] So MIT engineering, obviously like deeply technical, you know, your stuff and you, you, you, you know, you code, right? You build software. Um, and so I've heard a lot about how AI has completely changed engineering in software. And so I'm wondering if you could tell us how that has changed things. And then I, you know, maybe that will explain how it's going to change accounting.

Jessica McKellar: [00:43:31] Yeah, yeah. So. Here's my, I mean, I'm just honest take on this. Yeah. So I mean, this is the biggest change to software engineering as a profession probably ever. The, the way, the, the way that things have been done for the entirety of this discipline of. Sort of artisanally crafting and reviewing and deploying code is it's going away and it's going away quickly. And honestly, there's a lot of emotion for a lot of people tied up in this. There are a lot of people who are in mourning for the old way of doing things, that they have invested their careers and being good at going away. And maybe it doesn't go away everywhere all at the same time. And there are some kinds of engineering that are going to retain that artisanal feel. But on the whole, it's not even that it is changing. It's like it has changed and it's been a fast change. And so like, what have I observed about this? This transition can be really hard and is like deserving of a lot of empathy, compassion, understanding, supporting folks in the transition. And it's a transition to work that is fundamentally pretty different because it's not about producing the code anymore. It's about having good taste, good judgment, knowing what to build and like kind of how it should be built or architected.

Jessica McKellar: [00:45:11] Having good taste during review. A strong sense of empathy and understanding of the end user of the system. And there are some people for whom it's really exciting that that gets to be the focus area now. And then there are some people for whom that is not exciting, and it is a source of deep existential stress. I think there are like complete mirrors to basically every other industry, including accounting. And yeah, I mean, I think that, I think that if you are a person who loves bookkeeping or accounting or numbers and like being behind the scenes and, and like focusing on just being really great at that production work, I think it's a time of tremendous change and adaptation right now that is like deserving of a lot of support and empathy. I think that just being realistic, If that work can be done in software, it's often going to be. And then there's going to be some really delicious judgment strategy, client support work that's going to be on the table as an opportunity now and then. It's not going to be for everyone. And we're just all going to have to move through that.

Blake Oliver, CPA: [00:46:30] What part of accounting do you believe AI will not do?

Jessica McKellar: [00:46:35] Uh, the part that's being a therapist for your clients. I mean, there's a lot of like, look, I'm a business. I have these priorities. I'm trying to grow. I'm distressed and like, what I need is someone who I know I like and who's like walking the block with me, figuring out how we're going to get through this. And I think that's a very durable relationship. The other thing that I was okay. Two other things that I would say. Number one, in some of these cases, it's not that. It's not that the work couldn't be done a different way. It's just that people want to be like you. You're accountable for this. You, Sally a person with a face who I can call on the phone. You are accountable. And so if something goes wrong, I'm going to call you. And we're going to it's going to be on you to fix it. People want the accountability with a person and that that's durable. And then the other, just to state an obvious thing, like some of this stuff is credentialed, like tech. You have to be credentialed tax preparer to file someone's taxes. And like, that's not going to go away either.

Blake Oliver, CPA: [00:47:39] What do you disagree with when it comes to the consensus view of how accounting is going to get? Uh, it's going to change because of AI. What where are you?

Speaker 3: [00:47:50] What what is what is the consensus view?

Jessica McKellar: [00:47:52] Can you do you feel like there is one?

Blake Oliver, CPA: [00:47:54] Well, okay, let's think about it. Um, one of the views is like, uh, AI is not going to take our jobs, right? We hear this from, from thought leaders, uh, talking around at conferences and stuff, right? Like, don't worry about it. It's not going to take your job. That's one view. Another view is like, oh, it's going to make accounting amazing because we're going to automate all the tedious work and we're going to get to finally do the advisory work that we were born to do. So those are kind of two, two statements or arguments that I hear frequently. Sure. Do you agree with those? Do you disagree?

Jessica McKellar: [00:48:28] Yeah. Tell me if you think this is a fair answer. So first, if I constrain this to accounting as a profession for a second. I, I mean, I think it is clear to many people, including me, that a lot of the production work is, is can be done, can be delegated to AI. Like, so there are people who are like sort of still, I see on LinkedIn people who are like, AI is all smoke and mirrors. And I'm like, I don't, I just.

Blake Oliver, CPA: [00:49:00] That's another view.

Jessica McKellar: [00:49:01] From my experience, right? Like I just, I've like seen too much of this, like in production at scale for a long time. Like I just I don't agree with that. I think that it means that the work will change. I think that there's I, but I think within accounting, there's like a lot of opportunity to do different kinds of work, provide different kinds of value. There's like the finance and accounting spectrum, like from the bookkeeping through to the FP and A through to the CFO advisory. And just like we can play on that spectrum more with all this time we've gotten back for people who want to do that. That's what I would say within accounting. Like I would be lying if I didn't say that I, I think that this, this is changing all professions at the same time in a way that is unprecedented in human history. And I don't totally know what happens because of that, because I think that the result like this is like, not to my mind, analogous to like, oh, someone invented the, the cargo container and now like the number of stevedores at the docks, is like smaller. And so that's like a major transitional challenge for folks who are career stevedores. But like other kinds of jobs opened up. I get that, but that's like, not really the analogy because this is happening in every profession simultaneously. And I don't know exactly how this plays out. I think that there are some ways that this plays out that are like pretty scary. Um, and I like, want to believe in a world where we're all invested in and are like holding our politicians accountable to helping us create a world where there's like a strong. Support net for making sure that everybody makes it through a rapidly changing economy. So okay, so I don't know, that was like, is that an optimistic answer, a pessimistic answer or both? It's never been a better time to my mind, to be an accountant.

Blake Oliver, CPA: [00:51:00] Yeah.

Jessica McKellar: [00:51:01] I'm not exactly sure what's what's going to be happening for us all five years from now.

Blake Oliver, CPA: [00:51:06] It's a difficult question to answer because it depends. And I guess that's that's very fitting because that's often the answer when it comes to accounting, right? When you get a question from a client, it depends. I need more information. Yeah. And it seems like some people are thriving with AI. Like if you can harness it and use it and you have good judgment and experience, and so you can evaluate what it's doing and guide it. I mean, it just, it, it's made me insanely productive. I can't even put a number on it, but I'm going to say ten X. True. I'm a, I'm a ten X whatever you call me. Podcaster, startup CPE educator. Right? Yeah. Um, but then like, if you, if you don't, your job gets automated away. So it's like we have this, uh, divergence happening where some people are going to benefit a lot and already are. And then the folks who don't adapt, like, what are they going to do? Right? What are the what are the jobs that they're going to have? And we see this broadly across the economy, I think in a lot of places in white collar work. But, um, you know, in, in accounting in particular, it's, don't you, isn't it fascinating how this technological revolution is like the first one to really hit the white collar workers? Right? Because all the other ones have been like automating the jobs of blue collar Americans.

Jessica McKellar: [00:52:33] I think it's super interesting. Um, and will have important societal ramifications. And what, what, what a wild ride that we all get to be on together.

Blake Oliver, CPA: [00:52:46] I guess it's a really good thing that we've had this accounting talent shortage for so long, because it gives us a lot of slack, right? I mean, you know, the, the last I checked, right. It's so interesting that that, uh, when you look at all these charts of jobs or comparisons of the jobs that can be automated with AI versus the job losses we see in accounting. It's on the far right of these charts in terms of automated automate ability or the ability to be automated, right? Like all the studies say, our job is one of the, the easiest to fully automate or almost fully automate with AI. But then you look at actual job losses, like hiring of new graduates is a is a good metric and it's totally flat. We haven't lost. And other professions are like down 10% or more, right? 10 to 20%. And we haven't. And I wonder if that's something to do with the slack or the tight tightness in our labor market in accounting? Or is it something deeper where these analyzes of like our profession are saying it's easy to automate, but they're missing something? And actually that's, that's the last question I want to leave you with because you've been, you're not just a tech founder, right? You are leading an accounting firm. Right. And so what is your take on that? Like the, the resistance accounting has had to, uh, to job losses as a result of AI.

Speaker 3: [00:54:17] Yeah, I think.

Jessica McKellar: [00:54:18] It goes back to something we were talking about earlier, which is I think there's a very durable place in the economy and in just like the business connective tissue of our neighborhoods to have. A finance and accounting professional who really knows their stuff, who's like supporting business owners, I think that that's going to be really durable. And, and that's like not true of, of, of every, um, every specialization or profession this, but this really, I mean, I think basically every business owner benefits from having this expertise and the accountability, sitting with an expert who they can rely on and that, that, um, yeah, I think that that's really durable, that relationship.

Blake Oliver, CPA: [00:55:01] You know, that's the it's so interesting that you say that because that was like, that was the most impactful thing my mentor ever told me or taught me, which is it's all about relationships in accounting. So I guess that's, that's our, that's what's going to save us.

Jessica McKellar: [00:55:20] And I don't think that that's a bad like, you know, living in a world where as humans, relationships are more important than ever is like probably a good correction on what's been going on over the past decades, like less TikTok, more connection, I don't know. I mean, I don't know.

Blake Oliver, CPA: [00:55:40] Less, less doomscrolling more human connection.

Jessica McKellar: [00:55:42] I don't, I don't know, I don't know, I like I have two little kids and I like, think about this a lot in the context of raising little, little people and.

Blake Oliver, CPA: [00:55:51] What is the world going to be like?

Jessica McKellar: [00:55:53] Yeah. And I think like a world where it's like, actually, it's really important to us to have strong human connections because in a world where everything can be done by AI. That's fine, but people want to be connected to each other and have shared experiences like, yeah, I don't know. I don't know if there's something that is like beautiful in that.

Blake Oliver, CPA: [00:56:09] I think that's a great way to wrap things up. Before we go, Jessica, how can our listeners learn more about Meridian?

Jessica McKellar: [00:56:17] Oh my gosh. Sure. Go to meridian.com. Lots of info. I mean, look, I was literally doing some bookkeeping right before this call. And it is a joy. It is a joy. If it seems like any of what we've talked about might be helpful to you in solving your day to day problems or firm level problems like you should, you should come chat with us.

Blake Oliver, CPA: [00:56:43] Great chatting with you, Jessica. Hope to see you around soon.

Jessica McKellar: [00:56:46] Yeah. Thank you.