AI: Accounting Intelligence | AI Hype vs. Reality

Isaac Heller and guest Blake Oliver dive deep into the world of AI in accounting, examining recent VC investments and emerging trends. They discuss the potential for AI to augment and automate various accounting tasks, from bookkeeping to auditing, while exploring the challenges of implementation and data management. The conversation also touches on how these technological shifts may impact the accounting workforce and reshape the competitive landscape for both startups and established firms.

• (00:00) - Welcome to AI Accounting Intelligence
• (00:38) - AI and Accounting: The Hype and Reality
• (02:18) - Analyzing VC Perspectives
• (04:10) - The Death of the Big Four?
• (05:40) - Emergence's Matrix and Automation Potential
• (13:30) - AI's Role in Auditing
• (17:03) - Challenges and Opportunities in AI Adoption
• (25:22) - Future of AI in Accounting
• (29:52) - The Role of AI in Excel and Data Management
• (31:06) - Cost Constraints and Efficiency of AI
• (32:37) - Smart Query Structuring and Enterprise Search
• (36:14) - Big Four and AI Investments
• (38:21) - Challenges and Opportunities in AI Adoption
• (45:31) - The Talent Shortage
• (48:58) - Private Equity and Accounting Firm Buyouts
• (52:53) - The Future of AI in Accounting
• (57:08) - Conclusion and Final Thoughts

Learn more about Trullion

Connect with Blake Oliver
https://www.linkedin.com/in/blaketoliver

Connect with Isaac Heller
https://www.linkedin.com/in/isaacheller

Get NASBA Approved CPE or IRS Approved CE
Launch the course on Earmark to get free CPE/CE for listening to this episode.

Review
Leave a review on Apple Podcasts or Podchaser

Subscribe
Subscribe to the AI: Accounting Intelligence podcast in your favorite podcast app!

This podcast is a production of the Earmark Media

The full transcript for this episode is available by clicking on the Transcript tab at the top of this page

Creators & Guests

Host

Isaac Heller

CEO @ Trullion | Modern Accounting Technology

Guest

Blake Oliver

CPA and Founder of Earmark CPE — earn free CPE for listening to podcasts and watching YouTube videos

What is AI: Accounting Intelligence ?

Join Isaac Heller as he meets with leaders in the accounting technology space to discuss how AI and automation are transforming the accounting industry, how technology has evolved, and how AI can help accountants work more efficiently. You’ll also learn how accountants can embrace innovations to improve their careers and lives and get forward-thinking perspectives on where the accounting profession is headed when it comes to new technologies and AI!

There may be errors in spelling, grammar, and accuracy in this machine-generated transcript.

Isaac Heller: Hey, everyone, this is Isaac. I'm CEO at Trillian. And today on AI, accounting, intelligence. We'll be talking with real people about real things related to AI and the impact on the accounting industry. Stay tuned.

Isaac Heller: Hey, everyone, this is Isaac Heller coming at you with Accounting Intelligence podcast. I'm here with Blake [00:00:30] Oliver. Blake, it's really good to see you again. Isaac.

Blake Oliver: It's been too long. So glad we could, uh, collaborate on this.

Isaac Heller: It has been. You know, Blake, I feel like you and I were talking about AI and accounting. You know, 3 or 4 years ago at the lunch table in the corner. And now, you know, all the cool kids are saying, you know, hey, come, come sit with us. You know, we want we want to know more about AI and accounting. It's finally here, 2024.

Blake Oliver: Oh yeah. And not just the cool kids [00:01:00] like the big venture capital firms all put out blog posts about artificial intelligence, accounting, accounting, technology. Uh, for me, the fact that accounting tech and fintech and AI are all hitting. Like they're hitting big now with VCs is so exciting because I've been following this stuff for years and years and years. I mean, more than ten years, and now it's having its moment. And it must be so exciting [00:01:30] for you, as a founder of one of these companies, to to see the attention that accounting is finally getting in tech.

Isaac Heller: Yeah, definitely. Well, it's bittersweet because we talked to a lot of those firms at the beginning. We ended up getting great investors. But but now they might be, you know, provoking a few more companies to to start AI and accounting tech. So now we got to look over our shoulder.

Blake Oliver: Oh yeah. Competition.

Isaac Heller: I'll say if you're if you're listening now to me and Blake and you're, you're kind of dabbling in this AI [00:02:00] and accounting market. Whether you're a founder or an investor, I think you're going to get the good stuff today. If you stick around and, you know, we'll kind of, you know, Blake, I think we should we should take people through some of these high level trends and maybe even get into some, some predictions for the future in this market. That sounds.

Blake Oliver: Good. Yeah, yeah. So I think the plan of the day is right. We'll go through these three blog posts, the the articles by Andreessen Horowitz and Greylock and Mcap [00:02:30] and uh, we'll, we'll look at what they get right from our perspective. We'll look at what they get wrong. I did find a few interesting oddities in these theses that they have put together about the future. And, and it should be good because, you know, you've got the accounting and the tech background. I'm a CPA, I have a podcast that covers accounting. So like we're going to we're going to be the domain experts here analyzing [00:03:00] these these VC.

Isaac Heller: Absolutely. Well, we got we got you all covered. Well, let me let me bring everyone up to speed. Um, and you can go check it out. I'm sure we'll provide some links, but Andreessen Horowitz A16z put out their AI and accounting map. Greylock. You know, Greylock, the famous VC. They put out an article about vertical AI and one of the big verticals was professional services, including accounting and then emergence. Famous famous [00:03:30] B2B SaaS investor on the West Coast put out their article about all the different professional services that are going to be automated, and obviously bookkeeping and financial reporting are on that list. I think before I get your your thoughts on them, I just noticed that two out of the three have the word death in the title. You know, Andreessen Horowitz, something like death, taxes and AI. And then emergence is is death to the big four. So it's a it's a bit it's a bit [00:04:00] intense. Um, you know, I don't know if I don't know if death is the right word, but you know what? What caught your eye when you started reading up on on how these firms are viewing the market.

Blake Oliver: So the death of the big four, that headline, uh, on emergencies article caught my attention because the big four are huge. They brought in $200 billion in revenue last year, the big four alone. So we're talking about Deloitte, PwC, EY and KPMG. [00:04:30] And that's a lot of money. And so if you are a tech company and you just have to grab a tiny little sliver of that in order to be very valuable, right, to be a success. And so, you know, if in theory, the big four could be replaced by AI, that's a that's a really big, um, that's a really big argument to make. Right And we can discuss that. Ah the ah the big accounting firms, consulting [00:05:00] firms going to get replaced by artificial intelligence applications. Or is it going to be a different approach where it's it's they they purchase these applications and use them in their practices? I think those are the kind of the two competing ideas. And there'll probably be examples of both. Right. Because these firms do a lot. They don't just do one thing, they they do a million things. Um, can we start with that post? I just think that was the most interesting to me because they really [00:05:30] focus on that.

Isaac Heller: Absolutely.

Isaac Heller: And a little a little side note, the article was actually called Deaf to deaf to Deloitte when it came out. You know, a little alliteration. And I, and I believe they may have, uh, ruffled the feathers of perhaps an audit or advisory partner, which reminds you that, you know, these are very, I think, important and, and really beneficial brands in our ecosystem, whether or not they're using the right mix of technology and bodies [00:06:00] is is another discussion. But yeah, I mean, look, the the death to the big four essentially created a matrix, a map of a lot of different services activities. Right. So, you know, on on one end, maybe they, they argued that bookkeeping and certain financial reporting, uh, aspects were highly repeatable and highly automatable. On the other end of the spectrum, arguing that things like international tax and M&A diligence might be harder because [00:06:30] they're so specialized. And if you look at this matrix, it kind of goes through the different types of professional services and what could become more automatable. Um, what I like about the way emergence did it is they really focused on the problems, right? And the problem set and they got very specific about all the different services and adjacencies and map them out in terms of what could actually be applied AI and what might be better in terms of specialization. I don't I [00:07:00] don't think people think a lot of that. I think they just fly 50,000ft above and say, oh my gosh, everything's manual. It must be, you know, prone to AI, but it's really not like that. You know, AI is applied to every single professional services line. And then within the professional services line, whether it's revenue or leases or quality of earnings, it's applied very specifically. So I think they I think they got that right. And I think they did a good job.

Blake Oliver: Yeah I like this chart. It's really helpful to [00:07:30] look at and um, to help our listeners visualize it. It's a it's a classic four quadrant chart. And the x axis is is basically whether or not the firm is helping you do something like advising you on something or doing it for you. So over to the right as you move to the right. Those are those are jobs where the professional services firm or the accounting firm or the consulting firm is doing that work for you, for your company. You're just outsourcing it. And then the y axis, the up [00:08:00] and down it goes from, like low repeatability up to high repeatability. And what what emergence is saying is that they want their investments, their, their, their startups to focus on high repeatability tasks that can be done over and over again. Um, and where the firm is doing it for the client. So it's an outsourced activity that is repeatable. And to me, this screams, uh, productized service. That's the [00:08:30] phrase that we use a lot in accounting services. So if a service can be productized, it can probably be automated with technology. Or there's more opportunity now. Yeah. Yeah.

Isaac Heller: So what's what's interesting is, you know, and Greylock. The article on Greylock is called vertical I, and emergence is kind of unpacking. And if you understand the way investment theses work, a vertical in [00:09:00] most cases is a specific vertical. So, for example, a public company called Veeva Fantastically Successful Company, really focuses on life sciences and pharmaceutical. Right? A company called Procore, another public SaaS company, really focuses on the construction industry. And so there aren't a lot of big companies that just focus on accounting. I mean, we can name a few Caseware and, you know, some of the, the Thomson Reuters and CCH applications, but that's not as big as a lot of these vertical markets. So the argument [00:09:30] is, you know, these companies today, these services firms rely on high human intelligence. Right. And the next era is artificial intelligence. So what aspects can be A what aspects. So creating a whole new vertical. It's a whole new vertical SaaS market According to these predictions.

Blake Oliver: Do you consider like public accounting to be a vertical? Because public accounting firms serve so many verticals? [00:10:00] Are they themselves a vertical?

Isaac Heller: Yeah. I mean, in SaaS terms, everything is a vertical. You know, maintenance plumbing is a vertical. Right. And so the question is, is is it a big enough vertical for venture capital investors who previously had exits in Airbnb and Uber and Instagram? Is it big enough for them to have those outcomes? I would say public accounting is not big enough, right? Um, professional services is big enough, right? That [00:10:30] $200 billion figure you mentioned, a lot of it is from audit, a lot of it is from tax. And they have legal arms and strategy and transaction services and stuff like that. So professional services as a whole, from the big four to Accenture to big BPO shops, that's a huge vertical public accounting. It's big. But you know when there's when there's four players that dominate it, it can only be so big.

Blake Oliver: Okay. So within the public accounting vertical, emergence [00:11:00] calls out certain services as ripe for disruption or automation with AI. Bookkeeping is obviously on there. Financial reporting. Uh cloud migration. Yeah. Quality assurance it just maintaining apps, I find that's interesting. That's actually one of the most, um, repeatable, uh, tasks that they identify is just maintaining apps. Do you, do you have an idea to what what what they mean by that when they say maintaining [00:11:30] apps.

Isaac Heller: They I don't know, I mean, maybe it's DevOps, maybe it's orchestration. I have no idea.

Blake Oliver: Um, I'm wondering if it's, um, maintaining like API integrations or something because. Right, that's, that's a, that's a big task. And maybe I could like Do that? Absolutely.

Isaac Heller: Absolutely. I mean, again, in theory, you're looking across an organization and you're saying, okay, who who does something that's just maintaining a bunch of stuff [00:12:00] with very clear, you know, inputs, outputs, rules. And yeah, you might look at someone in the IT org or the CIO org and they say our job is to make sure nothing breaks in these APIs. And then you say, okay, well that's a service, right. And so we can build something that automates that service. Now I want to I want to throw something out there. I mean, we're talking about the venture capital community, which I can speak deeply about. Um, ambitious agenda, sometimes fatal flaws. Right. [00:12:30] All the ups and downs of going through, you know, the venture capital route, um, you know, remember, these are firms and these are industries where we put a lot of hope in blockchain. Right. And, and, and crypto and NFTs. Right. So I'm not saying we're wrong, but sometimes you you build the hype and the market first, right? That hype could be because you are raising money from LPs to show them that these, you know, [00:13:00] AI is the next frontier. Or that could be to encourage entrepreneurs to build. Um, or it could just be to stay relevant in a category that exists. So I think it's always important for us, you know, you and me, Blake, and really the practitioners to form our own hypotheses, you know, to take this in is that this is what could be this is what could be. But ultimately, we've we've seen these hype cycles go, go up and down. And this is probably a good example of one that's coming. [00:13:30]

Blake Oliver: Well, there's a part of me that thinks that emergence might be on to something here with this chart. They've got, um, the one thing that really strikes me about it that I find interesting is where they put audit. Auditing is right on the center line up to the top. So that means it's highly repeatable. But emergence is saying that it's sort of like in between A help you do it and do it for you. Kind of, uh, right kind of job. And I would actually say that most audits, at [00:14:00] least financial audits, are really on the do it for you side of things where you've got a lot of companies, like most companies, they buy an audit because they have to. It just has to get done. And it gets done the same way every single year. And so I have been wondering and theorizing about how odd it could be disrupted by artificial intelligence. You know, um, it definitely I mean, we have we have we have sayings in, in the audit world, like Sally, [00:14:30] same as last year, is an acronym. So if something is Sally, to me that means it's automatable by AI, right? Because what better use of AI than to feed it a prior year work paper and tell it what you want done for this year and give it the current year's data set. Could it follow last year? Probably. It doesn't require original thinking.

Isaac Heller: Right?

Isaac Heller: Well, you know. You know, first hand. [00:15:00] Yeah. It's a it's a good question. And I, I also should give a special shout out to people starting to care more and more about audit. Right. It was always behind the curtain. Right. If you're a VC, you don't think about it, right? Maybe when your companies get mature, you kind of check the box. But it's only it's only when you start to look behind that curtain and say, hey, Blake, what are you doing? Oh, well, we're just really we're asking for last year's financials to start our audit. Okay. And you start to say, oh, that's automatable [00:15:30] that's repeatable. And so I think people yeah, I mean, you know, emergence is looking at audit as highly repeatable. But it's funny because audit itself is in a it's in a it's inversely correlated with startups, meaning startups literally don't care about audit, right? Seed a b companies like. Don't slow me down. Do not tell me my R number is wrong. Okay. Um, you know that that [00:16:00] just passed the smell test for my valuation. And so, um, it's just funny. It's just ironic that now there's all these, there's going to be all these companies being built where, you know, you and I have talked about some of these early stage companies that are startups doing audit, which is really for the kind of the big large companies that are publicly listed especially.

Blake Oliver: It's a $60 billion market in the US. So I feel like somebody's going to come along and tackle that. We already see it with Soc2 audits, [00:16:30] these data security audits. There's startups that are doing it for a fraction of the cost of a traditional CPA firm, and arguably Giving a better customer experience. So I feel like it's inevitable that we're going to start to see them. Those tech companies creep into financial statement audits. Absolutely. I would not be surprised.

Isaac Heller: Well, wink wink. You know, trillions doing it. You can go check that out another time. But I [00:17:00] guess I'll I'll ask you, Blake, you know, you talked about these companies like, have you seen any recent rounds, whether it's early stage companies or companies or growth stage, whether it's rounds announced or, you know, we were I just saw you at the AICPA, uh, a few weeks ago. Like, have you seen some companies venture backed or private equity backed companies that have caught your eye?

Blake Oliver: Yeah, we were at the AICPA conference together. And of course, a lot of these startups are there exhibiting. I [00:17:30] actually didn't go into the expo hall this year. I was yeah, I didn't I didn't make it in. Um, I was a little too busy by the, uh, the pool and the cabanas. You know, that's become like the side, the side event. And I like I like the sun these days. Uh, but I feel like a lot of them, a lot of what we see, you know, we've got these startups, um, you know, Field Guide is a great example of an early stage startup in audit, but they're really more like checklists. [00:18:00] They're workflow software, and they're adding AI into these apps. But it's really for the auditor to use, whether that's an internal auditor or an external auditor. Um, same thing with apps like Floqast. I used to work at Floqast. It started as a closed checklist workflow software. Now they're adding in AI features to automate the work. Now, Floqast is designed and marketed to corporate comptrollers and CFOs, [00:18:30] not the auditors. But essentially it's doing the same kind of work, right? Where a field guide is reviewing work papers. Floqast might be creating them. It's it's it's two sides of the same coin. Right. Um, but we haven't seen anyone. I haven't seen anyone. Anyway, that is like automating it without the auditors. At this point, I don't know. Have you what? Have you seen Isaac?

Isaac Heller: Well, I'm. I'm with you. I mean, [00:19:00] nothing new under the sun, right? Ai is the next kind of checkpoint on the road to automation. And so you talk about field guide and carbon and canopy and all these, you know, audit workflow solutions. They deliver audit workflows. And that automation is desperately needed, uh, amongst uh, audit and services firms. Uh, the AI features, that's who they are, their features. You know, they're just kind of adding on things. You know, we mentioned, you know, when [00:19:30] you think about the AICPA, first of all, it's it's amazing to see so many new logos. It used to be a bunch of like Like R&D tax credit, you know, services companies and stuff. And now, you know, we mentioned True Wind, which is a Y Combinator backed, uh, you know, audit workflow or accounting workflow solution materia. Right. Ex meta, you know, data science coming into the space. Um, there's a lot of these, these new Y Combinator venture backed companies coming in. [00:20:00] Um, have I seen anything? It's all interesting to me, right? It's all interesting. And and when you when you think about the future, you hear the word agents, right? Agents that can kind of do the tasks for you. You can kind of. It's almost like. Yes, the analogy is, but it's almost like a like a little robot, I guess, to do it for you or something. And, and that, that will come and that will evolve. But that is such a big behavioral shift that I think [00:20:30] and, you know, you and I know a lot of auditors and accounting firms. Right. Many of them are still in SharePoint, in Excel, and so I don't see them going to agents tomorrow. I see them going to some automated workflows and then the agents kind of taking them slowly along for the ride, you know. But that's a really good point. We're still waiting.

Blake Oliver: But that's a really good point. Isaac. It's a lot of these big firms have just recently moved from desktop [00:21:00] to cloud, and they still have another. They still have a lot more room to move into more advanced technology. Ai is like a huge leap beyond where they are now. And so, I mean, you know, we haven't we haven't even seen any of these early or these later stage startups go public yet, right? I mean, we saw Audit Board got acquired. I guess that kind of counts these days because there aren't that many companies going public, right? That was like a $3 billion deal. But that's the first one that I remember [00:21:30] hearing about in this space. So that to me indicates that there's still a lot of room to grow for these companies that, like most enterprise companies, are not actually using modern cloud based workflow software yet.

Isaac Heller: Yeah, absolutely. And let's like, let's let's break that down a little bit. Okay. So we we just talked about the early Y Combinator stage. We talked about the VCs. And you know you got to think about realistic paths for these these companies. And I caution [00:22:00] the practitioners on the line who are listening of using early stage firms. Right. It's okay. It's okay to be early adopters with early stage technology companies. Obviously, I'm a big proponent of that because when we started truly in a lot of firms were like that. But in the playbook of venture capital, eight out of ten of their companies fail and two are successful. So a lot of these companies are not going to be able to get to scale. Whether that means finding product market [00:22:30] fit or actually building a large customer book that allows them to reinvest in the next generations and features of their platform. So that's that's a word to the wise cautionary tale. Now, in order to to actually materialize that investment or make a return on that investment, they need to be audit board, right? You mentioned audit board. Audit board was acquired last quarter for $3 billion by HG capital. That's roughly a 15 times return or revenue [00:23:00] multiple which is better than many public companies right. So that's.

Blake Oliver: Like twice as.

Blake Oliver: Much.

Isaac Heller: As a lot.

Isaac Heller: Exactly, exactly. And you mentioned, you know, the IPO market one stream just filed their S-1 yesterday. One stream is a CPM kind of closed management solution. So again now we're we're shifting a little bit into the office of the CFO. Whereas Audit Board sells to internal audit teams and one stream sells to financial you know, financial reporting and fpna teams. But but that's kind of [00:23:30] the ultimate goal in this industry. Somewhere in between, there are cases of audit software companies like a Caseware, which coincidentally, coincidentally also was acquired by HG capital. Right. And, you know, CCH and Thomson Reuters have done some acquisitions. But, you know, the real I think big market is in the office of the CFO. We've yet to see if the audit market will be venture scale. It's still a great market and there's a lot [00:24:00] of opportunity for innovation. So I do have a question, you know, to you and to kind of the broader group is is this going to be a drive by market where there's a lot of VCs throwing into these early stage companies because ultimately they got to get to $1 billion outcome. And that's not a product of a good that's not a byproduct of a good product. It's actually more a byproduct of demand and market size. Right. And adoption cycles.

Blake Oliver: I mean, I'm with you on on the opportunity [00:24:30] Being in the office of the CFO, that's the customer you want to be selling to is corporate controllers, CFOs in in private industry. Public accounting is great. Don't get me wrong. It can be a great market, but it's so slow to sell to. The sales cycle is really long. These firms are very entrenched in how they do things, and a lot of that is the billable hour model does not incentivize innovation the same way that saving costs is incentivized in a private company, and just getting stuff done faster [00:25:00] in a, in a in a private company with fewer headcount is so important in a finance team. So I still think that, yes, the the opportunity is in serving that customer. And if I were starting a AI accounting startup today, I would be targeting controllers and CFOs for sure 100% or people on their teams who have specific functions. And I want to go back to that article that you brought up the the Greylock one. Yeah. [00:25:30] Vertical I their thesis statement because I think it you mentioned something in there that I completely I think they nailed it with which is this concept of AI agents. They're the only VC that mentioned this in their article. They said that today the dominant paradigm is a human paired with an AI copilot. Humans do most of the work, and AI copilots augment our capabilities. But [00:26:00] in the next few years, they expect to see more examples of the reverse in a flipped model. Ai agents will perform the majority of the work, and humans will check and edit the outputs. So in a corporate accounting environment, that's very appealing because it's getting very difficult to hire accountants. There aren't enough of them, there aren't enough professionals in general. And that is going to get worse and worse over the next few decades. So who's going to do this grunt [00:26:30] work? I agents?

Isaac Heller: Yeah, I mean that's that's the hope, right? You know, you you can start to conceptualize this when you think about RPA, right. If you've ever implemented robotic process automation, you show them a, you know, repetitive task. And then the robot should learn it and should repeat it. Um, AI agents are like robots with, you know, a thousand times more, [00:27:00] uh, computational power. Right? And that's just, again, this kind of, this kind of curve of the AI experience. Look, I think it's going to be a gradual transition into this. And when people talk about AI, it's not just about artificial intelligence and the power of machine learning and llms. It's actually about a new user interface, right? It's about how we will interact with software. Um, to do a stretch analogy, the way that mobile came in and changed the way we interacted [00:27:30] with screens on a mobile device is something you could potentially foresee with AI. Meaning instead of a UI where I click and filter whatever report I want to run for this period and this entity and, you know, this exclusion, I would just say I would type into a little box, hey, can you show me the report for this? Right. So that's step one. It's around [00:28:00] what you would call task automation, right. Really just interacting with the software and that executing on on what you like to do. And then eventually that copilot, let's call it that person or that bot that's automating your task will say, hey, that's a great question. But actually you should look at the Italian entity because the question you're looking for of whose spending has the most vendors over $100,000 is happening [00:28:30] in Italy, right? And so it's going to be proactive about its response. Right. And then you'll get to the world where that agent will actually come to you at the beginning of the day and say, hey, France looks good. But Italy is really, you know, out of whack right now. You should take a look. Now, I don't think that's as achievable as as people think it is. Right?

Isaac Heller: And I know it's a lot.

Blake Oliver: Harder than it sounds.

Isaac Heller: It's a lot harder.

Isaac Heller: And and the as I say, the problem of of yesterday is the problem of today. [00:29:00] And the problem of yesterday is that we couldn't get a complete data set. Right. Whether you're a corporate controller, you just don't have access to the data. Right. Or it's in silos or you're an auditor. I mean, that's like an auditor's number one thing. If I could just get the data right, let me do my job right. Half of it is is pulling all the different siloed files and stuff. In order to have these agents perform, you do need to be able to traverse different data sets, right? And if you rely on the ERP, well, more than 50% of the data is outside of [00:29:30] the ERP. So then you got to go structure that. So it's really it's a nice future. It's a cool futuristic idea. But the finance stack is like a freaking mess to begin with right.

Blake Oliver: Yeah much easier.

Blake Oliver: Said than done, right? Like like for example, um, that example you gave where you said, I need to create this custom report. Could an AI agent do that? Well, yes. In theory, I just type in my request in the AI agent goes and builds the report for me in Excel and pulls in all the data from all the different sources. [00:30:00] But think about who's doing that job in your organization right now. That is a very experienced Excel ninja on your team most likely. And do we really expect that an AI agent will be able to match that person's skill level to the point where that person is replaced? Probably not anytime soon, but maybe if you've got like five Excel ninjas, you will only have 1 or 2, because now, instead of coding all of these formulas themselves, they can just ask AI to [00:30:30] code it and then check the result. And you can do that now in ChatGPT or in Claude, or in perplexity if you want to do something in a Google sheet. Now, Gemini just got rolled out in my Google Workplace account, and now there's a sidebar and it'll create formulas. It'll just it'll just build the the spreadsheet for you. So that's really exciting to see. But that is still the copilot approach. It's not the full AI agent acting [00:31:00] autonomously. Um, so it may take a while right before we see that. And there's also the cost constraint, right? Because AI is very expensive, and I think the estimate on like a single AI query could be like $0.30 for like a really complex, you know, Claude, three query, right? I don't know what they've brought it down to, but it's not cheap, right? It's meaningful each one. And so to run a continuous AI agent you're basically running prompts [00:31:30] constantly. How much would that cost? Would it actually be a savings versus maybe offshoring it somewhere? There's going to be this cost benefit analysis, right? Isaac.

Isaac Heller: Yeah.

Isaac Heller: We we assume that software operates at these beautiful margins. Yeah. Because we went from on premise to cloud. But it doesn't it doesn't necessarily mean that's the case. So the you know, the the theory, the Moore's law says that [00:32:00] we'll get the cost down. But it it's a new era. I think there's some advantages in accounting because language processing is I mean, it's all expensive, but language processing is not as expensive as like image processing or video processing. Right? So those are much more compute intensive and relative to, you know, what people are trying to do. Accounting actually ends up a little bit more self-contained. Now these these clod [00:32:30] queries that go over, you know, assume it's combing half the internet.

Blake Oliver: Thousands of tokens. Right.

Isaac Heller: It's expensive. Yeah.

Isaac Heller: Those can be more expensive. But you know, then you then you get into how you structure these queries. So like at Trillian, like if our client is querying something, it's our job to set up the query to make sure they're not querying all their data. Like, maybe there's a way to understand initially how to spotlight that subject [00:33:00] or that entity, or that bucket of documents that they're looking for to answer the question. So there's going to be some smart companies at the application layer. And and if you look at companies, there's you know, we talked about a company named Gebbia that announced 130 million from Andreessen And Horowitz. Um, companies like glean, which does enterprise search as well, I think a lot. I don't know heavy as well, but I think what a lot of those companies are doing is they're trying to figure out how to make that search [00:33:30] more affordable. That seems to be like the next phase. Um, right. Because otherwise you're double screwed, right? You're a VC backed company that can't break out to $1 billion outcome. And guess what? You ain't operating profitably anytime soon because your compute.

Blake Oliver: Your Cogs.

Blake Oliver: Is insane. Because you're feeding the entire ERP data set into an LLM every time somebody makes a query. Right. Which doesn't absolutely doesn't work.

Isaac Heller: By the way.

Isaac Heller: Like I have to say, if you're listening to this right now, Blake just dropped [00:34:00] a Cogs. If you're listening to this right now and you don't know what cogs is, don't start an AI accounting startup, right?

Isaac Heller: There you go.

Isaac Heller: Make sure you know what Cogs is.

Blake Oliver: Good.

Blake Oliver: Go to perplexity. Go to perplexity and ask, what is cogs? It'll probably give you a great example of that and ask for it in SaaS. That's my new favorite research tool. Um, I'm, uh, I'm beyond Google now.

Isaac Heller: You're you're in.

Isaac Heller: Claude. You're Claude guy.

Blake Oliver: Uh, well, I'm. I use Claude for writing. I find it has [00:34:30] the best writing natural. Writing. And now you can create projects so you can give it examples of your writing. Yeah, and it'll write in your voice. But I'm using. I just started subscribing to. Perplexity I the pro model. And, um, it basically is like Claude because it's connected to. Claude, you can actually specify that you want it to use Claude 3.5 sonnet and it searches the web. So it's. Like Claude plus Google. It's insane. Yeah. So it's like like let's [00:35:00] say I hadn't heard of one of. These startups that you mentioned, like Gebbia. I just say, what does Hebia do? In perplexity. And it goes out and it runs a bunch of Google searches, and then it summarizes all the information for me. So it's like doing that job. I used to have to go click into all the top ten results and read every article. Right now I've got that agent or that copilot doing it.

Isaac Heller: That's awesome. Um, I'd be really impressed if they said heavy is doing what we do, but for funds and [00:35:30] private equity. So, you know, it got me thinking, though, you know, we're talking about all these AI applications. Um, and those are at the application layer. I mean, you mentioned Claude and Perplexity, which are deep AI tools, right? I think AI infrastructure, the Llms we've talked about, you talked about copilot, right, within Google or Microsoft. Right. And we talked about the big four, right. Those are those are all players that [00:36:00] are kind of in the game. Right? You could be an infrastructure LLM and then get really good at accounting. Right. You could be Google and say and Microsoft and say, hey, the world is going to live on spreadsheets. We're going to keep investing here. And then you could be KPMG right. And EY and everything and say hey, we're the big four. The clients trust us. So we're going to we're going to navigate that. I mean the big four are investing billions, billions of dollars in AI. [00:36:30] Um, do you have a perspective on, you know, if they're going to make a claim to this, this AI accounting, or are they just going to keep announcing investments or any of the big players? Are they players?

Blake Oliver: I am.

Blake Oliver: Deeply skeptical. Whenever one of the big four announce that they are spending $1 billion on anything, because I wonder, where does that number come from? Is that a cash expenditure, or is that simply.

Blake Oliver: We're going to give.

Isaac Heller: $1.

Isaac Heller: Billion.

Blake Oliver: Right? Are we going to [00:37:00] are we going to give. Does it mean. Does it mean that they're giving everyone like an hour a week on their time sheet to experiment on AI? I'm not sure if this is like noncash imaginary Whatever it is. But regardless, um, I don't think they're very good at it. Professional services firms are historically very, not very good at building technology solutions. Like name one app that has come out of [00:37:30] an accounting firm that has that has.

Isaac Heller: Yeah.

Blake Oliver: You changed our lives, right? Um, they're internally built. Tools tend to suck. So, I mean, I it's going to be the innovation is going to come from outside and get adopted by the big firms. That's always how it works. Right? Um, and they'll influence it for sure. Right. I'm not saying that they can't help, but I wouldn't expect, like, you know, Deloitte to become the, uh. I [00:38:00] wouldn't expect Deloitte to figure it. Well, unless Deloitte figures out how to build a superintelligent artificial intelligence, that just becomes Deloitte, then, you know, I don't think they're going to do anything.

Isaac Heller: I love it, if that makes sense.

Isaac Heller: Yeah, well, we're still on the docket. We're still on the doctor Evil thing. $1 million to build a clone of itself, right?

Blake Oliver: Well, and that's.

Blake Oliver: Why we don't have to worry. We actually brought up the reason why we don't have to worry about a superintelligent AI anytime soon, which is that the compute power or [00:38:30] the the electrical cost to run a superintelligence would be, I think the estimate is greater than the entire output of the United States right now, the electrical consumption of the United States.

Isaac Heller: Oh, wow.

Blake Oliver: So so that's why Sam Altman has invested in fusion companies, because he knows that the only way to achieve his objective of actually creating an artificial general intelligence, which means an AI that can think continuously, like you and me and have memory and all that. The only way to do [00:39:00] it is with massive amounts of power. Or he has to wait years and years and years and years until we figure out how to make llms that are as efficient as the human brain or close to it.

Isaac Heller: Oh wow.

Isaac Heller: Well, that's. Yeah. I mean, it does it does kind of atone for our earlier mention of outsourcing because when you when you compare, like, you know, the human cost to the robotic cost and then you go all the way to AGI, it's kind of nice to know that to recreate [00:39:30] Isaac or Blake, you know, it's the entire GDP of the United States. So I it's.

Isaac Heller: Insane.

Isaac Heller: This is peak value. This is peak value, right.

Blake Oliver: And the human.

Blake Oliver: This is this is the the the the classic, uh, fact that everyone learned when they watched the movie The Matrix that the human, the human body and it's mostly the brain uses like 20W of power a day or something. It's insanely small amount of power for what the brain is capable of [00:40:00] doing. So it's actually like a miracle what our minds are capable of. Uh, and it's funny, you know, that AI is so inefficient Um, but it also means there's a lot of opportunity to lower costs. And and we've seen that already, like new models come out that are half as expensive within months. So if that continues, it's like Moore's Law, but division instead of multiplication in terms of. Right. Yeah.

Isaac Heller: It's it's I still can't do division. [00:40:30] So I don't know how, but yeah, it's all right.

Blake Oliver: You won't have to soon. The, the AI will do it for you.

Isaac Heller: Absolutely.

Isaac Heller: But look I on the, on the big four point, um, I think you're I think you're right. Um, you know, it's not to say that the Big four is not capable, right? It's just so it's been done so many times over the years. Right? And [00:41:00] it's not just the big four. It's that classic like business 101 use case when they say, oh, well, you know, the oil company. Standard Oil was great at drilling oil, you know. And why did they fail? Because they moved into midstream pipes or whatever it is. Right. You are a services firm, right? You are elite at services and building brand and delivering an A+ project. Building technology is not your [00:41:30] core business, right? Right. So, and you just read so many business cases over and over again where people try to go into their core businesses and they don't find synergies. Now, that being said, you also mentioned M&A. And I think that's the other thing, just to kind of tie all these, you know, AI and accounting investment trends is there is some more why now than just AI, right. Ai is is exciting. But from an accounting perspective there [00:42:00] is a there's a shortage. You know, not as many people are excited about going into accounting. You also have a wave Of, I would say, macro pressure on large accounting firms to operate more efficiently, efficiently. Right. It's just a natural progression over the years. And you know, you saw E almost got broken up this year. There's a lot of people that come into the industry saying what. You can't audit can't do consulting, consulting can't do audit. That [00:42:30] seems wildly inefficient. Why do you even have the businesses combined? Right. You saw the largest private transaction of an accounting firm last quarter, with New Mountain Capital closing on Grant Thornton, who had also acquired Citrin Cooperman as well. Another top you know, 50 firm. So the amount of accountants. Right. In addition to I the actual amount of accountants and just this general operational efficiency push like, hey, [00:43:00] that whole throwing bodies and spreadsheets at the problem is not scalable. Right. You can keep rolling up regional.

Isaac Heller: Firms.

Isaac Heller: You want, but I'm looking at my software company. I'm looking at my, you know, even car companies. Right. And they're just operating much better. So, um, just to kind of like circle back in addition to I, you know, these guys are investing in companies should be looking because now's now's really the time. There's a big why now compelling event for these companies to look at more software. [00:43:30] Well, you've I mean you've felt it, you know, working at Floqast and now kind of getting to talk to a lot of people in the industry. Um, it feels like it's inevitable.

Blake Oliver: Yeah. You bring up a good point about the talent shortage. Again, this is why companies are not as interested in replacing headcount as they are in augmenting their existing employees, because we just don't have enough people to go around as it is. And that is why I [00:44:00] think Greylock has the right view on this. They say in their article that they recommend startups taking multiple product multiple revenue stream approach. So don't try to solve everything all at once. Come up with a solution that solves this particular pain point and sell that and then expand.

Isaac Heller: Well, it's also the only way. You have no other option. There's not there's not it's [00:44:30] not like a CRM or an ERP where there's just this well-defined hub and you can go out and build the next NetSuite or workday or whatever it is, like. You got to you got to be like, hey, uh, payroll audits. Cool. Can I start there? And then I'm going to do expense audits and then revenue audits, you know, or whatever it is.

Blake Oliver: Yeah, whatever it is. Right. That that that repeatable task that takes up time every month that somebody on the team has to do. [00:45:00] You can automate enough of that. You know, you can find the right price even if you're just doing bank recs. It might seem like not a big deal, but how much time is spent every single month across every team in accounting, doing routine bank reconciliations, somebody is going to figure out how to create a really easy tool to do that and sell that everywhere. And there'll probably be a bunch of different options. So I think Greylock. Greylock got it right there. Um, [00:45:30] you know, we haven't really talked about the Andreessen post that much yet. I like, um, they they're the ones who really called out the talent shortage that you mentioned. Isaac. You know, the 75% of CPAs could retire in the next ten years. That statistic. That is wild. It's true.

Isaac Heller: Uh, is that the death part, or am I being too cynical? Death, taxes, am I? No, I'm just kidding.

Blake Oliver: Yeah. Yeah, I think that is that is the death part, right? Because, you know, traditionally, um, accountants [00:46:00] don't don't live as long as everybody else once they retire, sadly.

Isaac Heller: Wait, is that a thing?

Blake Oliver: Yeah.

Blake Oliver: Yeah, yeah. I mean, um, accounting firm partners, like. And this is the case, I think, with.

Isaac Heller: Don't tell me that there's a different life expectancy for accounting accountants.

Blake Oliver: Yeah, yeah. Because, well, it's for for anyone where you don't have a hobby. Right. If your work is your life and you suddenly retire, you don't know what to do. And I think actually, like, psychologically that actually causes [00:46:30] you to die younger. Right. Like purpose keeps people going.

Isaac Heller: Oh, interesting.

Blake Oliver: Develop some hobbies, folks. If you're if your work is your life, either never stop working. Like don't retire right right. Or or develop some hobbies so that you, you have something to keep you going when you retire.

Isaac Heller: Well, look, I think I think there's a, you know, there's, um, a mirror that the industry is looking in right now. I think there's a lot of self-conscious people. I mean, these are all my peers and [00:47:00] our employees at Trulia is. And our friends and our partners. And there they are saying, you know, how do we iterate? How do we create the next level of this? Because there is some sense of like, you know, if my head is on a swivel between two screens, one is an invoice and one is Excel, and that's not going to be very fulfilling for me. So it's definitely something it feels. It doesn't feel like that when I talk to our partners and I talk to our and maybe we just get to work with the tech forward ones. I mean, we get [00:47:30] to work with a lot, truly. And I'm talking about like, we we get to work with the advisory firms that are kind of on the cutting edge. And we do work with the big four a lot, but we also work with those firms of like, you know, I was a Big Four senior manager, and then I went to go to, you know, Connor Group or Riveron or Embarq or cross country, you know, these types of firms. And a lot of those firms are much more tech forward. You know, they're much more enabled. And they're not just like on one project for two years, you know, they're switching [00:48:00] from clients and stuff like that.

Isaac Heller: But I didn't even realize that about like, kind of the life expectancy or let's just call it the career expectancy, because that does have an influence. So if these accounting or accountants are retired, you know, it's one of those it's one of those baby boomer kind of equations of looking at actual like large curves. And then now that makes sense around this trend with of more PE buyouts for accounting firms, because I would guess that there's a large [00:48:30] class of partners. And of course you have a legacy partner compensation model. If nobody's familiar with that, partners make a good amount of money in many firms, and a lot of the work more and more is pushed down to staffs and associates, and over time it just becomes harder to make partner. But partners continue to do well. And so as these get to retirement age, I'm just kind of thinking out loud. Blake, this is another reason that there's, you know, they're more open to a buyout. Right? So you've got the labor [00:49:00] shortage. It's all connected with kind of career spans, the labor shortage. And then also these partners ready to kind of buy out and bring in the governance, which is going to make them more efficient with technology. Yeah. So very interesting.

Blake Oliver: And those those private equity buyouts are an opportunity for tech companies to come in and fill this gap in labor. Because you have these partners all retiring and there aren't enough staff to do the work. And tech is going to have to fill this gap. Um, [00:49:30] so it's going to be interesting to see whether or not these investments are successful. I think it will all depend on how successfully they modernize these firms. It's essentially the same thing we see with small firms getting flipped and acquired. Um, and it can be very successful and very profitable. I can't even imagine how hard it will be to do in a large firm, given all the complexities and the number of partners and the number of people.

Isaac Heller: We are in it.

Isaac Heller: We are in them, and it is. It is very, very hard. But [00:50:00] that's what private equity specializes in, right? Like I've been you know, everyone you talk about it from a, you know from a 50,000 point view. Been in these rooms. You know, I've worked for PE backed companies. I've talked with the PE firms as they're talking to accounting firms. And they are very used to going and partnering with legacy companies that have ten, 20, 50 years of file folders and over time [00:50:30] being able to impact change. Right. And it starts gradual and hopefully you can get the right sponsors internally. But over time, once you start seeing those benchmarks and that governance, you know, the company just has to turn, it has to go more efficient. So, um, I don't think I don't think it's anything they haven't seen before. I mean, go, go look at a healthcare roll up or something like that. You know, they've they've seen a lot of paper. Let's just say that, um, so Blake.

Isaac Heller: Maybe some of those.

Blake Oliver: Medical.

Blake Oliver: Is a great example. Yeah, [00:51:00] some of those have gone really well and some have like not right. We've seen I think in the dental world. Right. There's been some roll ups where the dentists feel like, oh now I'm working more. I don't have a stake in this. Like and that's, that's where I see it could go wrong is if, if a private equity firm buys a big firm and then they just decide to, like, squeeze it as much as they can so they can sell it in 3 to 5 years. That would be bad for the staff because then they're just going to get asked to work a lot more. But then there's the other option, which [00:51:30] we hope they take, which is let's modernize, let's implement technology. And and that could lead to a really positive world for accountants in public, where they aren't being asked to work 50, 60, 70 hours a week anymore. And that's what I really hope for. That's the promise of technology, is that accountants can have good work life balance. It's totally possible. There are small areas in accounting where we do have that. I was privileged to have it and I hope everyone [00:52:00] can. So, you know, I really am passionate about tech because it can give us that really good job. Pays well. Stability. We don't have to work more than 40 hours a week and we can do it remotely. I mean, that's the dream.

Isaac Heller: You love.

Isaac Heller: You love the road. I'm totally with you.

Blake Oliver: I love.It.

Isaac Heller: I've worked for for under private equity firms. I know that there's been some historical negative things, but overall, um, I think I think driving efficiency [00:52:30] in a, in an industry that hasn't changed in a long time is probably the only alternative. We don't know how it's going to go, but the status quo certainly, um, you know, is going to lead to more articles with death in the title or, or some kind of negative or cynical connotation. So I'm totally with you. Um, Blake, you want to I want to end it on kind of a question or thought So we talked about these early stage companies. Right. We talked about the firms. [00:53:00] Um, how do how do these exits go in ten years? Right. Are these are these early stage companies IPO ING. Is it PE buyout. Are they disappearing? Are they combining with firms? What are you thinking?

Blake Oliver: Well, if they if they're just providing a small vertical solution, meaning like a task solution, they'll probably get rolled into something else at some point. Because long term, like long term, these little point solutions are going to [00:53:30] expand into basically the AI agents we talked about and filling the headcount role, like actually taking the place of, say, a staff accountant. Eventually that might take ten years or more to happen. We will have the choice of do I hire a human staff accountant or do I hire a AI staff accountant? That would be the winner of this, right? That company will go IPO and be worth a lot of money. Maybe [00:54:00] there will be a few of them that come out of this. So like that's the you know, to me that's the like driverless Uber car kind of outcome where the profits will just be insane because, you know, as long as you can keep the compute costs down, you're going to be ten times less expensive than a human. And so you can you can sell that all day long.

Isaac Heller: Yeah. Does that make sense?

Isaac Heller: Yeah, definitely. And and you know, a highly you know, [00:54:30] I would say human accounting advisory and services will just command a premium. Right. It'll be a premium service in the same way that that high end travel is concierge based. Or, you know, another analogy is, is diamonds, right. You can buy a fake diamond and nobody can tell the difference. Right But some people want a real diamond, right? And that's okay too. And so so there is a premium to working with someone and having that look and [00:55:00] feel and having that relationship. And then in some cases maybe your your rev rec or your lease audit or whatever the heck it is, you want the automated path, right? And that's okay too.

Isaac Heller: And oh yeah. Nobody's going.

Isaac Heller: Anywhere. We're all going to have some fun.

Blake Oliver: And it's going to start with jobs like payroll specialist and accounts payable specialist. Like, these are jobs that we don't have enough people for. And we can put those people to better use by upskilling them with AI to do other things. So [00:55:30] like, I'm not really worried about AI taking away jobs because there's just so many jobs to be had, right? We're barely keeping up with economic. We're not keeping up with economic growth as it is.

Isaac Heller: And so we need people.

Isaac Heller: We need more.

Isaac Heller: People. Wait.

Blake Oliver: But it's going to be crazy. It's going to be a wild world where your coworker is an AI, like in slack or in Microsoft Teams, you will have a bunch of coworkers who have names, probably because we love to humanize things. Who are AI agents that work with you [00:56:00] on your team? Your manager might be an AI. Actually, that's probably one of the easiest jobs to automate.

Blake Oliver: I.

Blake Oliver: Think is is just like the routine. The routine middle manager job is just following up to make sure stuff gets done.

Isaac Heller: I love it, just just show me my KPI. Remind me my KPIs once a week.

Blake Oliver: The AI Isaac I'm willing to bet that we will see companies run by AI CEOs [00:56:30] not in our lifetime. Meaning like the board will just delegate the management and direction of the business to an AI because these AIS are going to get really good.

Isaac Heller: Oh my gosh, I can tell you some days in my role I would love to have an agent who could take a day or two off.

Isaac Heller: Wow.

Blake Oliver: They can sit in on all of those. They can sit in all of those meetings and you can just do the fun. All hands. Right, Isaac. [00:57:00] You can go pump up the team.

Isaac Heller: All hands. And, you know, the AICPA talk or whatever it is. So I love it.

Blake Oliver: Yeah. There you go.

Isaac Heller: Yeah. Well, look, I mean, you know, Blake, this is this has been great. And we should, you know, we should do it again. We should check in every year and see how things are going. Um, and the AI and accounting world, but, um, I really I really appreciate your time and insights and, you know, enjoy you as a thought partner. And, um. Yeah, it's always been fun kicking it for a little bit on topic. [00:57:30]

Blake Oliver: Yeah. Thanks for having me on. This was a great conversation. We could go for hours.

Isaac Heller: Okay, good. So part two coming soon for all you listening, and we'll see you next time. Thanks, everyone.