Orchestrate all the Things - Trullion is applying AI to create financial workflows in accounting. Featuring Cofounder / CEO Isaac Heller

Episode Date: March 26, 2022

There's no business like powering business. That may seem like a paradox, and most of us don't really think of accounting as a separate line of business. In that sense, it's a secret hiding in plain sight: accounting is big business, in and of itself. The global accounting services market is expected to reach $735.94 billion in 2025, and digital transformation has targeted it for impact. The size of the accounting software market is projected to reach $22,9 billion by 2027. This includes names from Microsoft, Oracle and SAP to Intuit and Xero, plus many others. Yet, there's something none of these names are applying, says Isaac Heller: AI in accounting. That's why Heller and Amir Boldo co-founded Trullion in 2019. Trullion seems to be on a mission to disrupt the accounting software market by "unifying the unstructured and structured worlds of accounting by reading PDF or Excel-based contracts and translating them into financial workflows, connected to the data source". In February 2022, Trullion announced that it has closed $15 million in Series A funding. Trullion has more than 100 clients, a mixture of large enterprise clients and audit firms which procure directly for their clients, and has crossed the seven figure mark in recurring revenue. We caught up with Heller to discuss Trullion's focus and its use of AI to innovate in the accounting software market. Article published on VentureBeat

Welcome to the orchestrate all the things podcast. I'm George Amadiotis and we'll be connecting the dots together. There's no business like powering business. That may seem like a paradox and most of us don't really think of accounting as a separate line of business. In that sense, it's a secret hiding in plain sight. Accounting is big business in and of itself. The global accounting services market is expected to reach 736 billion in 2025, and digital transformation has targeted it for impact. The size of the accounting software market is projected to reach 23 billion by 2027. This includes names such as Microsoft, Oracle, and SAP, two Intuit and Xero, plus many others. Yet there's something none of these names are applying, says Isaac Heller, AI in accounting. That's why Heller and Amir Bolto co-founded Trullion in 2019. Trullion seems to be on a mission to disrupt the accounting software

market by unifying the unstructured and structured walls of accounting by reading PDF or Excel-based contracts and translating them into financial workflows connected to the data source. In February 2022, Trulian announced that it has closed the $15 million dollar series AI funding round and it has more than 100 clients, a mixture of large enterprise clients and auditor firms which procure directly for their clients. We caught up with Heller to discuss Trullion's focus and its use of AI to innovate in the accounting software market. I hope you will enjoy the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn,

and Facebook. Yeah, definitely. So, Twitter, LinkedIn, and Facebook. Yeah, definitely. So, George, yeah, great to talk to you. Glad we connected. You know, my name is Isaac Heller, and I'm one of the co-founders at Trulian. In my role, I act as CEO currently. By way of background, I grew up in Texas. And, you know, as it relates to Trulian, I lived in the world of finance and accounting. You know, I went to an MBA when I was in Texas and learned about Gap and I have these big standards. But then when I went into the business world, the corporate world, I ran into these really, really challenging new standards and new business models. Things like revenue recognition, revenue accounting, lease accounting, and such. And a few years ago, actually in New York, I got an advertisement for a company called Lemonade to sell me insurance. And I looked up this company,

Lemonade, and it was a fantastic company that was a global organization with R&D in Tel Aviv, right, over in Israel. So I had this idea of, you know, building an accounting platform, but doing it in a global way. And so with my, you know, wife and daughter, we moved to Tel Aviv to build this company, which is now Trulian. Along the way, I met my co-founder, Amir Boldo, and CTO. And we've been at it for many years now. So we're really excited that the market has validated our product seed, Series A, with some great investors. From a product perspective, we do something really cool with accounting. So we use AI to kind of automate the extraction of data from documents in Excel. And then we also do the financial entries and the workflows that are currently done outside of these big ERP systems.

What's cool about that is it connects the source data with the financial entries and what are called disclosures. And that allows both controllers and CFOs and their auditors to access everything in real time. So that's a little bit about us. That's a little bit about me. And happy to share more. Okay, great. Thanks for the introduction. And one thing that was not entirely clear to me just when doing some background check, let's say, on Trulion and what it does was what it is that the company focuses on precisely.

Initially, I had the impression that it only focuses on a specific, let's say, subdomain of accounting, which is called lease accounting workflows. And I got this impression because you seem to be putting out that message a lot. But then, and again, hearing you describe the company in your own words, I wasn't so sure. It also sounds like you don't just focus on this niche, let's say, but you also aim for the bigger domain. So which one is it? Would you be so kind as to clarify that? Yeah, it's the big domain, right? And it's been the big opportunity from day one. So I would think of it like a picture in your head of three parts that are disconnected.

The first part is, you know, the corporate accounting world, let's say, right, that might be doing some of their processes and spreadsheets or offline. Okay, the second part is the ERP, which is SAP, Oracle, and Workday, right? And then the third part is the audit leaders, so the audit team. So those are in three different parts. There's a lot of reasons why the ecosystem is getting more manual and riskier. We won't get into that. But the question is, is how could you make it a more accurate and transparent ecosystem? And the way to do that, in our view, would be to find ways to connect those three parts. Right. Meaning everyone in all parties could connect on one source of truth.

But in order to do that, if you try to come to market with a, you know, a heavy tool that did it all, it wouldn't be successful or you would have to spend $100 million just to start. So what we do is focus on a pain point. In this case, we focus on things like lease accounting and revenue recognition. And these are just one of the many workflows that these three different areas are working on, but they're workflows that have not been fully automated or fully connected. So the vision has always been the same but as you know george like you really have to focus and start in very practical ways in this market okay okay i see thanks thanks for the clarification i guess that explains the current focus of the company on actually what caught my eye, to be honest with you, was the mention of some specific standards, which I was entirely unaware of, as probably most people

who are not in accounting standards with exotic sounding names like ASC 842 and IFRS 16 and the like. And it seems like the reason why you are elaborating your messaging around those is because there's an imminent change. So this ASC 842, which is apparently a new standard for lease accounting workloads is about to come into practice right and so your your messaging is that well it's a good opportunity for for companies to update their accounting practices and Trillion is a good vehicle for them to do that yeah absolutely so if you think about the accounting world it's always you know the goal is always the same.

We want more transparency and more on the balance sheet. And so these standards, these accounting and revenue recognition are just the latest and greatest. Think about it like, for your listeners, GDPR, right? We all know what GDPR is and these new regulations, and they drive a lot of innovation. This is the same type of thing, but it's in the accounting space. is and these new regulations and they drive a lot of innovation. This is the same type of thing, but it's in the accounting space. Okay, thanks. And so again, following up on this thread of trying to position, let's say, what you

do in the broader landscape. Obviously, even to someone who's not familiar with the domain, such as myself, I only have like a passing acquaintance, let's say, with what's going on there. But I do know that there are solutions that are aimed at people working in that space, some of them like Xero or QuickBooks. And then there's even some more specialized ones that focus on this specific niche on lease accounting. So solutions like Re-List or Visualiz. So how would you say that Trulion is different? So what's the differentiation between those specialized and less specialized solutions? Yeah, definitely.

So with QuickBooks and Xero and the ERPs like NetSuite, we're complementary, right? So lease accounting would be a gap in their ecosystem. And so we fill that gap. And to your point, there's other lease vendors that do that. However, they don't have the AI and the OCR component that we have. So it's really a revolutionary idea. I mean, everyone always thought that you could just buy the solution and put it in manually, but we're the first person to connect all the work that's being done outside of it with the system itself. So ultimately it's a much better value proposition,

more accurate and lower cost of ownership for the company. Okay. Well, I have to say that as far as I've been able to investigate, what you say seems to be the case. So I haven't been able to identify an approach, a technical approach similar to yours, which you can confirm or disconfirm that, but to my understanding basically comes down to applying AI techniques, let's say, to ingest and integrate data from various sources. And to me or to anyone with technical background let's say this this seems kind of obvious and I was very surprised to see that it doesn't seem to have been applied much if if at all why would you say is that well first is that the case indeed then if yes why would you

say that is so it's it's definitely the case i think that's one of the reasons where you know where we are as a business and we're backed by these great customers and great investors um look accounting in general is underserved right uh there's a lot of startups out there that do you know sales software and marketing software customer support software and those are use cases that we all have experienced in early stage companies, but accounting, that's a big business. That's sort of the mature companies. So in general, it's not as well represented with startups. I think that there are some use cases where AI is applied, obviously, like accounts payable and invoice automation that are more common.

You know, for us, I don't really know, George. You know, don't spoil the secret. But we just hadn't seen a lot of AI. We saw a lot of people talking about it, right? People wrote some really nice papers about it, but we didn't see it practically applied. It's definitely harder than everyone thinks so even with us you know ai is not waving this magic wand and giving you a hundred percent of a document extracted but from an accounting perspective what it's doing is if you can connect that ai and that ocr to the actual financial workflow and let's say it gets you 70 improvement at least you can connect the document with the workflow.

So instead of being in Dropbox or another system, it's connected with the workflow. And so what that does is it makes the whole world more accurate because what auditors do on the other side is they sample. I don't know if a lot of your listeners know that, but the audit world, which is massive, that's the big four, right? And many others, what they're doing is doing a sampling process of one in a thousand transactions or one in a hundred contracts to produce a financial audit. And so if you can connect more of the source data in a really clean way with the financial workflow, using AI as one tool, like it's super, super powerful for this industry. So I don't know why more people aren't doing it, but it's definitely an exciting time.

Yeah, and since you mentioned the big four, and to anyone not familiar, that refers to the big four consulting firms. That was also another kind of naive maybe question I had. And I was wondering, isn't what you do something that those consulting firms and others apparently do probably regularly as part of their projects and is it that have you seen that and is it that you basically came out with came up with the idea of making a product out of it while they do it perhaps in a more ad hoc manner?

Well, to some extent. I mean, we love the big four, right? These are our partners. You can look. We have national partnerships with different big four. We have one of the big four that actually audits us from a stock compliance perspective. So we work closely with them. But yeah, I think everyone would agree in

the big four that they want to move more and more to technology the question is do they have the dna to do it right so there might be a lot of good initiatives for technology within the big four they might have big it teams but if you ask um you know the the young auditor at these companies if they're really producing technology that they're proud of or that they could sell, I don't know the answer, right? So we enter into a dynamic where a lot of the big four and these other accounting firms are realizing that their DNA and their specialty is services and accounting and technical accounting. And they're partnering with us as a software of choice for these new standards.

So we really believe that's the way the market goes, that we end up being a really, really strong partner for them. And that's what's happening so far in our business. OK, so let's shift gears and start talking about technology a little bit because obviously that's very, very important in what you do. You talked earlier about how a big part of what Trulion does is actually getting access to different sources for the kind of documents and information and data that you need to ingest. So the first and actually that makes a lot of sense. That's the first thing that anyone that wants to integrate data for whatever reason needs to do. You need to connect to where that data lives.

So I was wondering if you can share with us a little bit what kind of sources that's truly connect to and what kind of techniques are applied for each source. As far as I've been able to tell, it seems like some of these sources are files really, so nothing too complex or sophisticated such as Excel files and text documents, PDF or others. And there's different approaches for ingesting data from those, but I was wondering if there's also other sources that you connect to and how do you approach that? Yeah, definitely.

Well, first of all, I mean, it's called source-based accounting. that you connect to and how do you approach that? Yeah, definitely. Well, first of all, I mean, it's called source-based accounting. So this is something we've talked a lot about. And a lot of times when you rely on the ERP or an old database as a source of truth, it doesn't have any of the source within it. So with source-based accounting,

we're not just thinking about the PDF. Okay. By the way, as a side note, PDF, just to be clear, we have embedded OCR. So that's the first layer is normalizing these documents. And then we have NLP, which is natural language processing as a layer on top of that. Okay. So that's just the document side. But we want to be able to provide this for any data source and be a single source of truth.

So that extends to things like Excel, okay? And it's not just uploading in Excel, it's actually us being able to use different methods to translate the different data in the Excel and make it easy. And it's also a source system. So if it's a Salesforce or or quickbooks you know we're an api friendly system to congest the data so there's nothing there's nothing really super

special besides the ocr and ai but at the end of the day it's um normalizing all these different sources into one you know source based accounting system okay so in order to do that, as you hinted at, you need, I mean, in order to create one source of truth, let's say, or one repository, whatever it is you want to call it, you need to have a technical infrastructure. So I was wondering if you can share a little bit on how you do that. So to use a database or some object storage system in the cloud or anything else. And then you also need, I guess, a little bit of schema, a little bit of domain knowledge actually translated into schema so that you can define the main entities that you want to deal

with relationships between them and their properties and so on so anything you are able to share on those would be much appreciated yeah for sure well first of all george it sounds like you've either interviewed a lot of tech companies or you have some experience yourself. I do actually, it's the latter. Okay, very nice. Well, it's super impressive and I'm sure we could learn something from you, but yeah, you're right. If you take the accounting world, for example, it's dozens of different workflows, right? From a tax workflow to a lease workflow to a revenue workflow. And then there's even sub workflows within that for different verticals. And within each of those verticals, there is kind of an entity model,

as you mentioned, of the type of data you want to extract. So we do aim for those types of structures that allows us to continue to improve our, you know, not just OCR and AI, but also the general normalization of all the different data sources. But then at the same time, accounting is very specialized, right? So if you ask anyone, there's so many special cases and unique areas for every different workflow. So we give the system a degree of flexibility so that our accounting partners and audit firms can configure it in their own way.

You already mentioned the fact that you have proprietary OCR. So it seems like you have developed your own way of, well, how can we put it simplistically, identifying let's say handwriting on scanned documents basically. And that kind of stood out for me for two reasons. First, because I would imagine that most of the documents, most of the sources that you have to ingest are actually not handwritten. By now, I would imagine that most of them are actually, most of the data you have to ingest either lives in some database or worst case scenario, some Excel file, or even worse, worst case scenario, some document somewhere. I wasn't imagining that you would have to ingest

handwritten documents much. And then to add to that, again, I'm not the world's leading expert on OCR, but I was kind of assuming that that's a solved problem, more or less. OCR has been around for ages, there's tons of research and tons of software that has been developed around that. So why the need for you to develop proprietary OCR? Yeah, it's a good question. So first of all, we don't see handwriting as much anymore. We're not trying to solve that. But look, OCR to some extent is a commodity.

But what happens is you start with the commodity product, let's say Tesseract, right, or whatever it is. And then you start to realize that there's still a lot of improvements for your specific industry, right? So take what they call noise, for example. So splotches on a document or things being darker than lighter. And so you get into these things and you need to solve for all the fringe cases. So we don't try to reinvent the wheel, but we take a pretty good wheel and try to improve it a bit from an OCR perspective. I can't say that we're putting all of our expertise into improving OCR right now. It's more just to help our clients get the data they need.

And that's really it. AI is a component here. It's a component of the bigger source-based accounting world and accounting workflow. Okay, well, yeah, that makes sense, actually, because I would be very much impressed if you said something like, you know, it's a cornerstone of what we do. What seems to be more central to what you do, however, is named entity recognition. And that's basically what happens when, well, either through OCR or through a natively digital document or whatever way

you end up with some kind of text then in that text you have to identify, well, entities basically, entities of interest whether that's specific concepts that are of interest in your domain or even specific names such as the name of the CEO of some CEO you're interested in or some company you're interested in or whatever and in that reading one of your white papers again there was something that stood out to me and it seems like kind of quoting from the white paper here that you're combining deep learning with ontologies and this is very much state of the art I have to say and you know still under research and I don't think I have seen it applied in production much. So I was wondering if you could share a few words on who developed it, when, and how does it work in a nutshell?

I can't get too deep. I can say that Amir Boldo is our CTO. He's fantastic. He's built amazing teams and platforms. Shlomo Agostin on our team is the machine learning lead. And these guys are the ones behind the seeds. I didn't know it was state-of-the-art, but I'm not surprised when you told me. So that's all I can say.

We have an amazing team and they love what they do. Okay. All right. So then I also had one more technical question, but I think, well, it's probably one for your CTO as well. So I'll skip that and let's move to the last part of the conversation, since we also have just a few minutes left. So initially you mentioned that you just secured 15 million Series A. And I was wondering, well, first of all, congratulations, of course. But I was also wondering, so in order to get that,

obviously the investors must have liked what they saw. So could you share a little bit of what you shared with them, with the rest of us? So what's the commercial traction like? Any client names that you're able to share or verticals or any other company fundamentals? Yeah, definitely. Over 100 clients to date, a mixture of large enterprise clients and also audit firms which procure directly for their clients we have you know a little more than half of our businesses in the US and a little more than a little less than half is outside so international you know obviously you got across seven figures

and recurring revenue to get to that stage so you know we've had tremendous growth not just in the past year but but in the past few quarters. And we've got amazing investors. So Aleph and Graycroft led our seed round. At Aleph, general partner Michael Eisenberg led the round. He experienced, you know, WeWork and Lemonade and Melio and these fast growth companies and started to understand what it meant to be accounting and audit ready. And then, you know, obviously when it came time for the series A, there were a lot of different options, but we went with a third point, which is Rob Schwartz and Dan Loeb, who know because they're a hedge fund, they know a thing or two about financial transparency. So yeah, you know, we have amazing people at the table, both from an Israeli investor side,

but also really a US and global investor side. Okay. So it sounds like the broader goal, the end game for you is basically to disrupt accounting software, but what about your intermediate goal? So what's on your roadmap for the next, let's say, a year or two? Yeah, definitely. I don't know if it's disruption. I think it's innovation. I think it's being a partner for the industry.

But certainly along the way, there's a little bit of disruption. We got to hire. We have to hire hire engineers we have to hire our sales org you know a lot of them are in the United States and some overseas and we we really haven't spent much in marketing and so we're going to start to bring our brand and our product really to the broader international market we signed a new lease in central Tel Aviv, which is really fun. And obviously, we've got the US teams as well. So that's it.

We're having fun. Okay. Okay. Sounds good. A couple of key figures actually I forgot to ask you about. So A, the year the company was founded and B, current headcount since you mentioned you intend to grow the team. Yeah, 2019 and we're at about 30 people right now.

Okay. Okay. And I guess what's your goal actually? How large would you like your team to be in a year from now, let's say? In a year from now, we're going to pass 50, 5-0 people, and it could go a lot faster based on how things are going for us.

Okay. Great. Well, thanks. Well, best congratulations for the funding again, and well, good luck with everything going forward. And thanks again for your time. I hope you enjoyed the podcast. If you like my work you can follow Link Data Orchestration on Twitter, LinkedIn and Facebook.