Why Accounts Receivable Is Fintech’s Biggest Untapped Market With Caitlin Leksana, CEO of Fazeshift

Accounts payable has produced multiple billion-dollar companies, yet its mirror image, accounts receivable, remains almost entirely manual at most enterprises despite decades of software spend. In this episode, Caitlin Leksana, co-founder and CEO of Fazeshift, explains why AR stayed unsolved and how her company’s AI agents are changing that. A mechanical engineer turned BCG consultant turned founder, Caitlin came to the problem the hard way, doing her own AR by hand at a previous startup, and her outsider’s view of a stubborn back-office chore is exactly what makes the conversation worth your time.

What We Covered

• A million AR analysts doing manual work in the US
• Why accounts payable got solved and AR did not
• The leverage imbalance between AP and AR departments
• The swivel chair problem and fragmented data
• $200 million in unapplied cash on one balance sheet
• Fazeshift as a context layer, not a rip-and-replace
• Why traditional SaaS and if-then logic could never scale AR
• The collections, cash application, and AR inbox modules
• Human in the loop and building trust when AI touches money
• Training agents on historical data and tribal knowledge
• From Y Combinator to a Series A led by F-Prime
• The vision for the context layer and autonomous finance

Key Takeaways

• AR is the inverse of AP, and every bill is someone else’s invoice, so the market is at least as large and mostly uncaptured.
• The real unlock is not the AI model but unifying fragmented data across the ERP, bank, CRM, and inbox into a single context layer.
• Human in the loop with full auditability is what earns risk-averse finance teams’ trust, and it is how agents move toward full automation over time.
• Some of the best unsolved startup problems are the ones furthest removed from an engineer, because no one with the tools to fix them ever felt the pain.

About Caitlin Leksana

Caitlin Leksana is the co-founder and CEO of Fazeshift, a San Francisco startup building AI agents for accounts receivable. She earned bachelor’s and master’s degrees in mechanical engineering from Georgia Tech, advised Fortune 500 companies at BCG, and earned her MBA at Harvard Business School before founding a crypto marketing startup and then Fazeshift. The company went through Y Combinator’s Summer 2024 batch, raised a $4M seed led by Gradient Ventures, and announced a Series A led by F-Prime in 2026.

Cleaned Transcript

Caitlin (00:10): There are a million AR analysts in the United States alone, which is equivalent to the number of public school teachers there are in this country. You just envision all public school teachers. That same number of people are actually just doing accounts receivable, everything from cash application, billing, collections, responding to incoming customer questions, calling people up. And so I think the scale of the problem is something that a lot of people don’t realize.

Peter (00:36): Today on the show, I’m delighted to welcome Caitlin Leksana, the co-founder and CEO of Fazeshift, a company building AI agents for accounts receivable. Now, Caitlin studied mechanical engineering at Georgia Tech, but then went to BCG to do management consulting. She then earned her MBA at Harvard Business School and co-founded a crypto startup before landing on the very personal pain point that became Fazeshift. In our conversation, we discuss why accounts receivable has remained such an unsolved problem, even as accounts payable has produced multiple billion-dollar companies. We get into the swivel chair problem that keeps AR teams buried in fragmented data, how Fazeshift acts as a context layer on top of existing systems rather than ripping and replacing them, and why a human-in-the-loop approach is essential when AI touches money. We also cover their modules across collections, cash application, and the AR inbox, and Caitlin ends with her vision for owning a wide-open market. Now let’s get on with the show.

Peter (01:54): Welcome to the podcast, Caitlin.

Caitlin (01:56): Thanks for having me.

Peter (01:57): My pleasure. So let’s kick it off by giving the listeners a little bit of background about yourself. Why don’t you just take us through what you’ve done in your career to date before Fazeshift.

Caitlin (02:10): So, hi everyone, my name’s Caitlin Leksana. I’m the co-founder and CEO here at Fazeshift. Before Fazeshift, I actually did another startup in the crypto space, we’ll come back to that. But for now, did another startup, and then before startups, I was at Harvard Business School where I got my MBA. Before that, I was at BCG doing management consulting, specializing in the transportation sector and then also in pricing, which was my specialty in the Atlanta office. Before that, actually majored in mechanical engineering from Georgia Tech. So I thought I was going to design motorcycles and race cars. That was kind of where my passion lied. Did mechanical engineering. Actually ended up starting a PhD program, dropping out, landing at BCG, figuring out more the business side of things, going to business school, meeting my now co-founder and CTO. And it was interesting because technically I was sponsored by BCG, meaning in consulting when you get your MBA, you’re supposed to go back to consulting, there’s a track to become partner, it’s a very sort of traditional career trajectory. And then I met my co-founder and CTO who convinced me that startups are a lot more fun than consulting. And here we are.

Peter (03:14): Okay, well it depends on the person obviously, but I would tend to agree with you. It’s interesting, like you started as a mechanical engineer, how do you go from focusing on that to focusing on accounts receivable automation? It’s not a logical path.

Caitlin (03:32): It’s not. I think the one, well, maybe two threads, but definitely one is, I like to solve hard problems. And I think that is the thread that has been consistent throughout my career. So in mechanical engineering, what you’re doing is you’re trying to solve a problem. And then through BCG, similar, you’re working with a business to try to solve that problem. Usually it’s with the CFO of a Fortune 500 company. Then the startups, like really good startups come from finding a really unique problem that nobody has figured out before. So that thread has been consistent. The other thread is just trying to make people’s lives easier, which is really what we do at Fazeshift. Throughout mechanical engineering, throughout consulting, it’s trying to solve hard problems and make people’s lives better. So that’s the thread. To speak a little bit more to the actual journey itself. So what happened was I went to Harvard Business School during COVID. During your MBA, you kind of have a lot of conversations about your career. That’s where I ended up meeting my co-founder and CTO. We were actually in the same section at Harvard Business School. He wanted to do a startup and so we just started brainstorming startup ideas together and we landed on this crypto idea. So it was kind of like a MarTech marketing community software for crypto, which at the time in 2022 was a really hot space to be in. And to give you a little bit of context of my co-founder and CTO, he was MIT computer science undergrad and masters. He then spent seven years as a nuclear submarine officer in the U.S. Navy and then taught cybersecurity at the U.S. Naval Academy before going to Harvard Business School. So that is his background, but he has sort of an expertise in technology and cyber. And so we went into crypto, kind of started ideating from his perspective, landed on this idea, and then we decided to do that together. So we raised our pre-seed round in 2022, right as we were graduating business school, did that startup. And it was interesting because we had customers and we had traction, but we were first-time founders and it just didn’t feel like the market was pulling us. We just did not have product-market fit, and crypto collapsed and all of that. And I remember one day I was going through the spreadsheet that I kept, which was the spreadsheet of all of our customers, how much we were supposed to invoice them. It was color-coded based on whether they’d paid or not. And it was just this realization for me that it was like one of the most painful parts of running this company was doing our own accounts receivable. So that is how we got to this position. It was a very personal pain point. It was a workflow we’ve had to do ourselves manually. And by that, I mean, I was managing this process. It’s customer-facing, so it’s sensitive. It touches money, so it’s sensitive. Interestingly enough, in crypto, we had our wallet ready to receive money and everyone wanted to pay us in fiat, bank to bank. They didn’t want to pay us in crypto. So we had actually a very, very traditional accounts receivable problem. And I’m an outsider. Like, you know my background, you know my co-founder. We are not finance and accounting people by trade. And so we took this approach that’s like, I’m an outsider just trying to learn more. So I knew I had the pain point. I took that to the market and I cold-called and cold-emailed probably a hundred different VPs, directors of finance and accounting, talked to accounting firms, I talked to CFOs, I talked to controllers. And even after a cold outreach, they would respond and be willing to get on a 30-minute call with me. And I would not lead the witness, I would just say, can you tell me about your AR process? Like, how does it work? I wasn’t asking anything specific. I was like, I have nothing to sell. I’m trying to learn, just tell me about this. And they would get on a 30-minute call and just rant at me for 30 minutes about how painful AR is and why it’s so painful and what their team does and all of that. And then even if they didn’t get on a 30-minute call with me, they would still send a very thoughtful response to my cold outreach being like, here’s all the bulleted lists of things that make AR painful for our company. And if you solve this, you’re really onto something.

Peter (07:17): So it’s interesting to me because AR has been around for, well, centuries, shall we say. There’s many established companies that have claimed to have created efficient systems. I mean, you’ve got all the ERPs, you’ve got fintech companies that just focused on this. But to your point, it still remains for technology companies a major pain point. Why hasn’t it been solved, do you think, already?

Caitlin (07:42): Yeah, this is a main question that we asked ourselves when coming into this space. Once we decided to pivot to Fazeshift, actually a lot of people recommended us against it because of what you mentioned. It’s crowded, other people have tried to solve this. It’s like, there’s nothing unique in the space. But because we had felt the pain point ourselves, we knew that something was still unsolved. And to give you some numbers, there are a million AR analysts in the United States alone, which is equivalent to the number of public school teachers there are in this country. You just envision all public school teachers. That same number of people are actually just doing accounts receivable, everything from cash application, billing, collections, responding to incoming customer questions, calling people up. And so I think the scale of the problem is something that a lot of people don’t realize. So that number helps put in perspective. Now, when we think about sort of what the market could be here, I look to accounts payable as sort of the best corollary comparison. If you look at the AP side, there is not only an incredibly crowded market, but multiple $5 billion-plus outcomes. So if you look at this from your investment TAM perspective, there’s still companies getting funded for AP and it’s still such a massive market that investors are still pouring money into it. And there’s multiple really, really good outcomes. There’s the old-school sort of Coupa, Tipalti, there’s the new-school Brex, Ramp, Zip, all these others. So it is not only a big market, it’s a big enough market that can sustain multiple huge venture outcomes. And if you believe that AR is really just the inverse of AP, meaning everybody’s bill is someone else’s invoice, you have to believe that AR is just a bigger market. It just hasn’t been captured. Like the visual I think in my head is like the Gartner Magic Quadrant report or something like that. AP has like hundreds of logos on it. AR maybe is really HighRadius, Billtrust versus paper. The question that we ask ourselves and that I think investors are also asking is, well, why? Why is AP so crowded and AR is not? And there’s two reasons for this that we believe, and it’s what we uniquely are solving for. One is who has the leverage in the relationship. The reality is that AP departments set the process and the AR departments conform to that process. And so the result is, if you’re in the AP department, you have one process that you set. It’s easy to set that up in a software. Sure, you have escalations and branching paths and whatever, but you get to define that process. On the AR side, you have N number of workflows that you have to set up per unique customer. An example of this is we have a customer that sells to Target. They have one dedicated AR headcount that is just dedicated to submitting invoices to Target’s portal in a very specific way or else Target won’t pay them. And Target has a different process, as does Whole Foods and Wegmans. And you kind of get how this explodes into a really massive amount of unique edge cases and nuances and configurations and complexities and back-and-forth email. That’s really where AI kind of helps us out here. And we started back in 2023. It was kind of right time, right place. Right as ChatGPT was coming out is when we really started to come into this space and realize what AI could unlock. So that’s one. The second, and this is what we think most other companies in the space have not solved for that we are uniquely solving for, and it’s actually nothing to do with AI. It’s what we call the swivel chair problem. So one of our customers coined this term, and he’s got a, they’re on HighRadius, which is an incumbent solution, and they have a 16-person cash app team, and they still have $200 million in unapplied cash backlog. Meaning $200 million on the balance sheet that they say, we do not know how to apply this cash because our team is so backlogged with cash application. And he’d said, my team is literally swivel-chairing between all of these different systems on their screen. So if you picture the visual, they have multiple monitors pulled up. They have NetSuite. Maybe they have Salesforce. Maybe they have JP Morgan or a bank pulled up for the bank statements and lockbox. They have multiple emails pulled up. They have spreadsheets pulled up. And his problem that he identified was it’s a swivel chair problem, which means fragmented data. So what we do is we are not a rip-and-replace. We’re not replacing the ERP. We’re not replacing the bank. We’re not replacing your credit card processor. We sit on top of those existing incumbent solutions that have been around for 10-plus years. And we’re automating what those 16 people are having to do every day and yet still having the unapplied cash backlog. So we’re solving the context problem. We become a context layer on top of all these systems. That’s what you need to do to properly solve AR that nobody’s really done. That plus AI allows us to come in and actually capture a market that’s completely greenfield and open for the taking.

Peter (12:15): That is really interesting because you started this before ChatGPT comes out, or as it was coming out. You go to your website right now and I’m on your homepage, “Your AI agents for accounts receivable.” So you are front and center with not just AI, but AI agents there. So tell us a little bit about how you’ve kind of morphed into this current offering.

Caitlin (12:38): We could not solve our clients’ problems without AI. And the reality is we actually use AI in a lot of different places. And it’s interesting because if you look at the market, people like HighRadius, BlackLine, they capture really large enterprises, but they’ve never been able to really dominate mid-market. And I think the question that we ask is, why can’t something like a HighRadius solve, or traditional software, like let’s not even put a name behind it, why can’t traditional software, traditional SaaS solve AR, and why hasn’t it today? But if you think about what I was talking about where you have N number of workflows per customer, the only way to solve that with traditional SaaS is actually just a ton of if-then statements, conditional logic statements, which we’ve seen people do in NetSuite. They will write hundreds of if-then statements in NetSuite to try to solve this problem. But that takes years of work. It does not scale. It is not a solution that can scale at SaaS margins. So that is really why we see traditional SaaS not be able to solve this. And when we were coming out, we knew there was a problem, we were trying to solve it. All of the prior softwares have done an incredible job with the software they’ve had in place. But now with AI, we can actually solve that way more than was solved historically. We can actually capture a market, we can do AR at scale, we can do things without writing hundreds of if-then statements. That’s what’s unlocked the market. I think we still realized there was a problem in the space. Without AI, we’d still be trying to solve it. It would just be much harder and the market would be much smaller. AI has given us the tools to be able to solve that at scale, which wasn’t before possible.

Peter (14:14): Right. And I imagine Target’s going to have their own way to pay people, as is every Fortune 500 company. In fact, every company of any size, particularly public companies, you go through the procurement process and then you have to follow these particular steps. Is that what your AI is kind of doing? It’s like looking at that 20-minute process and saying, it’s just a bunch of disparate data fields that it can probably just figure out itself?

Caitlin (14:40): That’s spot on. And like that process you just described, imagine that at 50,000 customers, which is what some of our customers have. They have like 50,000 customers. And before AI, the only way to solve that was to throw more bodies at the problem and say, you have to go into this portal and set it up and give them our information and make sure everything’s good to go. And then next customer comes on board and you just have to keep hiring and hiring and hiring. So we can take that sort of 20-minute process. We actually work with a bunch of different industries, but we had one HVAC company down in Florida. It used to take them three days every time they got paid a million dollars from one of their customers, because it was actually for a bunch of different invoices. We can use AI to break that down to less than a minute. And I think what this shift enables is these AR teams, instead of having to do this manual logging into portals, submitting invoices, checking payments, back-and-forth emails, is they get to, one, review the AI. So we’re moving more into a kind of a review phase, especially given the sensitivity of finance, but AI can really save them all that original time that they were spending. And then the second thing is they can be more strategic. So they can work on those accounts that maybe need more of a human touch or are a little bit more sensitive, or can look at the data that Fazeshift gives them and gives them more sort of decision-making power of like, hey, actually, if we changed our process to accept these sort of payments, that would really unlock a lot for the business. So I think the role is changing because AI can actually take a lot of this very manual work off their plate and save them some time. And then the other thing I’ll mention is, we were just talking about the AR versus AP comparison. AR is the process of how you get paid. That process you were just mentioning, you were trying to get paid the money that you were owed. I’m sure there was a contract already signed. Something was, you were just trying to get paid, and AR is the lifeblood of any business. It’s how you get paid by your customers. And it’s still a very unsolved problem compared to AP.

Peter (16:29): AR, it is making money. So you would think it would be actually just as, if not more, focus for companies. But it seems like we’ve just decided that this is the way it’s been and we have to put up with it. Is that what you’re finding in the marketplace?

Caitlin (16:44): I have a thesis which is some of the best, most unsolved startup ideas are those that are the most number of degrees of separation away from an engineer, because it means no engineer has ever had to solve it. And if you look at it, like, you know how many usage-based billing platforms there are out there? A ton, because engineers usually get tasked with building usage-based billing systems. Or even on the AP side, most kind of employees know, when I need a new tool, I have to go through procurement. I have to make sure that I get budget and then I have to bring this new vendor on board and vendor onboarding. And so tangentially, people have seen that problem. I think on the AR side, it’s just a really hard problem. And most engineers have never had to do cash application before. So I think coming at this from a more technical outsider who hasn’t been in the fintech or accounting or finance space, but brings a very tech-forward approach and having gone through YC, we get access to some of the most cutting-edge technologies and infrastructure. And we’ve had to rebuild our own infrastructure multiple times on the AI front because of how fast the technology is moving. So you’re somewhat marrying this like, technology’s moving fast, my CTO understands here’s how we can leverage technology, but then having a ton of empathy for what these people are doing. And if you look at the historical, there’ve been a lot of AR solutions. I think AR needs a rebrand at this point, but there’s AR solutions that are really credit card processors. And yes, they’re letting customers pay easier, but they’re not solving the crux of what makes AR hard, which is this fragmented data, back-and-forth, edge cases, nuances that differ by industry across company. No one’s really solved for that yet. And I think that’s really where the biggest opportunity lies.

Peter (18:29): Right, so we’ve talked a little bit about the invoicing side and the reconciliation you have to do. What else does a Fazeshift AI agent do?

Caitlin (18:39): Of course. Are you familiar with like Zendesk and Intercom and some of these? Yeah. So support teams have had tools probably for 10-plus years. And what those tools help is incoming requests, incoming emails, assigning it to the right person, triaging it, getting responses back, tracking it, all that good stuff. AR teams have similar volume, if maybe not even more, emails, requests, questions coming into their AR inbox, and yet have never had a tool like that. So that’s basically one of our modules, is helping AR teams use AI on their AR inbox or their collector’s inbox to triage, tag, understand what the request is, and then draft responses to those incoming questions to save them a ton of time. Most people say they cannot get to all of the emails coming into their AR inbox, which if you think of that from a support standpoint, a leader in the support org would say that’s ridiculous that we’re not responding to all customer requests. But the AR team just doesn’t have the capacity to. So we give them AI tooling within their AR inbox that allows them to actually respond in a timely fashion with all of the data. So our agent will go into the ERP and grab the statement. It’ll grab invoice PDFs. It will grab a W9 and make sure that’s all attached. Make sure we have all the information from the contract and that the terms are there, draft a response so the AR team can get it out faster. So that’s a key module that like nobody’s searching for that. Nobody knows that exists. But when we marry what can technology do and what is the core pain point that these teams are trying to solve, the more we understand how they work and what technology can do, we bring those together and then we bring those new modules out to our customers.

Peter (20:21): So then does that mean you are sort of agnostic as to the ERP, CRM systems that your companies are using?

Caitlin (20:29): We are completely agnostic to any of the underlying systems already in place. So any ERP, CRM, whatever, Outlook, Gmail, portals they may be logging into, payment processors. I think the thinking about agents is really thinking about how can we come in to their existing systems that are in place and help these teams automate faster and at scale, not ripping or replacing existing software. And part of our deployments are figuring out, okay, what are the teams doing? Are they logging into Target’s portals? Are they having to, I don’t know, send a bunch of emails, like whatever it is, and then we bake that into our agent so that our agent can take that on.

Peter (21:09): So then let’s say you’re a corporate out there, you might be using NetSuite. What’s involved in bringing on Fazeshift to help with all the problems they have?

Caitlin (21:21): So really at the crux of it is all the integrations. That’s always where we start. AI is a commodity. It’s democratized. Everybody’s using the same models. It’s either OpenAI or Anthropic’s. There is no reason to build your own model these days. Those models are just getting so good so fast. So if everyone’s using the same LLMs, then really the only differentiator is the data that you feed in. So AI is only as good as the data you feed in, which then comes back to, well, how do we get all the data? And that goes to the integration. So we always start with integrations to the different systems. Then we work with our team to figure out, what are the processes? What are the edge cases? What is your team doing today? How can we bake that in? We also train it on historical data. So our agent can go into the ERP and say, hey, you’ve reconciled these hundred payments this specific way. We know next time we see a payment from XYZ, we’re going to post it this specific way. So we can train on historical data and then we’re really trying to learn the tribal knowledge of what the team is doing today. And then we go into a human-in-the-loop mode. So I think this is really unique to Fazeshift. It’s what we learned very early on from dealing with finance and accounting teams, is you have to build trust in the AI. That’s why I believe there’s some startups that are taking a fully agentic approach to this where there is no human, there is no platform they can log into to see what the AI is doing. We’ve taken the complete opposite, which is we give you full auditability, full transparency. Your human team can see everything that the AI wants to do. It’s kind of like if you’ve ever used Cursor or any of those AI agentic coding tools, they’ll give you a plan. AI will say, here’s what I want to do. I want to go into this file. I want to make this change. I want to add this function. It’s the same thing that our agent does. And then a human person can come in and hit approve. If they approve it, then our agent will go and take all those tasks. There is an audit log, a history. And then if they make changes, our agent’s constantly learning from that. And then from there, we kind of take their team through that adoption curve where they kind of understand, here’s why AI is making the decisions, here’s how it works. Once they get comfortable, we can move more and more to full automation. So it’s a very, I think of it more of a consultative approach than traditional software where it’s like, in their hands and they just take it. We really tried to work with the team because the reality is every company and every team does AR differently, and a pure plug-and-play software clearly hasn’t solved AR, so we have to take a little bit more of a hands-on approach.

Peter (23:42): And I imagine then that the AI learns as it goes and it gets better with time. Is that correct?

Caitlin (23:47): That’s exactly it. So the AI is fully self-learning, which means even though to these AR teams, it might look very much just like a traditional software system that they log in, they see what the AI wants to do, they hit approve, they hit buttons, they add comments, all that good stuff. But as they use the system, it’s actually learning how the human teams would be doing this. Let’s say an edge case comes up, they can make a suggestion. The AI might already make a suggestion and they can just hit confirmed. Otherwise, they can kind of tweak things via the platform. Then they hit approve, then that reinforces. It’s a reinforcement technique that the human teams are basically training the agent, just like you train a new employee. Here’s how I handle this case, here’s how I handle this case, to eventually being able to handle over 90, 95% of the cases that come in.

Peter (24:29): So we haven’t talked about collections yet. Is that a part of what you guys offer? And obviously there’s plenty of companies out there just focusing on that one piece. Is that part of your offering?

Caitlin (24:38): It is. So we’re a multi-module platform. We have collections, cash application, AR inbox, credit, portal management, payment portals for collecting credit cards, all that. The thing with all of these different modules is they all make each other more powerful. So to give you maybe an example, let’s say I submitted an invoice to Target. Target then disputes that. Someone needs to see that in the portal, or else someone’s going to start sending a bunch of collection emails that don’t realize that actually there was a dispute in the portal. So you somewhat need all of these things to be talking to run an effective sort of AR strategy. So we do do collections. It’s one of our biggest modules, tends to be one of the first things we deploy because there’s very, very quick time to value. The nice thing about deploying AI within AR is it’s very measurable. So we can see how many dollars were collected, reduced DSO. The other thing about AI agents in AR is, yes, it’s measurable, but it also hits on more than just efficiency. When we’re talking about AI agents, it tends to be, how do you save time? How do you save money? How do you make it more efficient? But actually, if you can do collections properly at scale, meaning completely personalized with context, collection outreach at scale, then you actually get better collections, reduced DSO, less write-offs, all these things that AR teams are strategically trying to impact. And then they can go to leadership and to the board with all of these success stories. And we usually see that in the first month or so of deployment.

Peter (26:08): So is that typically the wedge that you go out with then?

Caitlin (26:11): Surprisingly, cash app is the biggest. Like people will come up to us and say, cash application is the bane of my existence. We went to Sage’s future conference and people will just come up and say, can you please solve my cash application problems? So that actually tends to be where people get the most excited. But then yes, we usually also sell collections and then the AR inbox, which is any sort of incoming email or text message. AI will read it, understand the context, route it to the right person, draft a response, grab the right documents and queue it up for an AR analyst to send back out. Those are usually the three that we sell in tandem. They’re the most common, they’re the most sort of quick to get out. And then things like credit management and portal automation are usually the cherry on top, but it depends. Every industry is different. We have worked with a lot of logistics and transportation companies that have hundreds of portals they have to log into. So think like Target times a hundred. And they have just people that are dedicated to logging into these portals like once a day, once an hour to figure out what’s going on. But we can do that with AI now.

Peter (27:08): So then do you have metrics, I know you’re still a pretty young company, but do you have metrics on what difference this has made through some of your customers, like hard data?

Caitlin (27:19): Oh yeah, we have a ton of case studies. It depends on sort of like, we actually build ROI cases for each customer that comes through our sales process. We very much want to make sure that it’s a good fit in both directions. So we ask them for things like, what is your DSO, AR per aging bucket, how do people pay you, how many people do you have on the team? Once we get a lot of that information, we actually put an ROI together. And honestly, if the ROI is not big enough, we’ll say, I don’t think we’re going to be a good fit. We can’t make much impact here. You should look into some other things. But if it is a good fit, we usually have multiple million in ROI for a small company. But yeah, we have a client. It was one of our first clients that had a nine-person team doing cash application every day. So they would log into JP Morgan. They would export their bank statements. They would export their lockbox into Excel rows, and they had nine people that were just going through that, reconciling which invoices those were for, and going to the email and then back to the Excel, and they had nine people coordinating on one Excel, but they had to get it done by the end of the day, or else they would get yelled at, because they need to have that back into the system so that it’s properly reflected. You can’t really do collections until you finish cash app, because then you’re just sending emails for incorrect data. You have to finish the application. And so they were spending all day, nine people. We cut that down to two people basically spending an hour a day, just handling exceptions. And those exceptions were defined by them. They were like, hey, we just want an extra touchpoint. We want humans to kind of keep an eye on this. These specific cases, we built that into the prompt. So the agent hands those off to the team, and we’ve gotten better with time. It starts with like 90%, I think we’re up to 95% full automation. We post those directly back to the ERP and no human takes a look at them. That’s how we’ve gotten them through the AI adoption process. They just fully trust Fazeshift.

Peter (29:03): Right, so when did you land your first client?

Caitlin (29:06): So we went through YC summer of 2024, and that was when we landed our first clients. We grew from zero to $100K ARR in two months, raised a $4 million seed round led by Gradient, which was Google’s dedicated AI fund, grew the team, just announced our Series A led by F-Prime, and are continuing to get an incredible amount of traction. My sales team continues to say this feels different than any other company they’ve been at. Once you get someone to a demo, the product just sells itself, which I think speaks to, if you really understand what these teams are doing every day and you build a product around that, then it really resonates when you show them. It’s the little things. It’s like, when a payment comes in, you don’t have remittance in the email. Fazeshift will automatically email that client, letting them know, we got a payment, can you please send the remittance? Once the remittance comes in, we automatically post that. And it’s like, sometimes just those little things that make people realize like, we understand how hard and how complex this job is and we’re actually building solutions for you.

Peter (30:03): So you’re still a very young company. What’s your vision though? Where would you like to be in, say, three years’ time?

Caitlin (30:10): We are targeting AR, and AR is our biggest focus, and it’s going to be for quite a long time. The metric that I mentioned of there are a million AR analysts in the United States alone, you can kind of do your own TAM calculation, but it’s a completely greenfield market that nobody’s captured. AR by itself is a big enough market to become a $1 billion, $5 billion company. So we have no doubt about that. Our main priority for the next few years is to dominate AR, and there’s plenty of pain points to go after. But the real crux of what we are as Fazeshift is the context layer. So we sit on top of all of these systems. It’s why we always hit IT and procurement, is because we integrate with some of these businesses’ most sensitive systems. Their source of truth, their system of records, their emails, their CRM, their ERP, their bank data. And we bring all this into one place and then we unify it, which means Fazeshift knows customer A in Salesforce equals customer whatever, X in NetSuite. That unification of, I know this customer in all context, I know the last time they emailed me, I know the last time they sent a payment and this is what it looked like, I know who the sales rep is that sold this deal, I know all past payment history in NetSuite. All of this in one place, this is the context layer. That is what is valuable in the world of AI, is the data and the context and making sure that you have that correct across all of these different systems. So wherever we go next, I think there’s a lot of really seamless places to go. Our customers at the end of the day are going to be the ones that drive this. People come back to us and they ask us for AP automation and they ask us for treasury and all these things. And it’s usually, you guys have all the data, you have the integrations, your team has helped us solve AR completely. Can you do that for this other process we have? So I think with that context layer, that’s really the vision that allows us to go any direction after AR.

Peter (31:55): Well, we’ll have to leave it there, Caitlin. That was so interesting chatting with you today and great to see that the AR side of the house is being attacked, and attacked with a vengeance, I would say. It’s such a waste of time for so many of the companies, and it’s great that we’ll be able to become much more efficient from now on. So anyway, best of luck to you and thanks for coming on the show.

Caitlin (32:17): Thanks for having me.

Peter (32:24): It is intuitively obvious that every bill is someone else’s invoice. But we have spent a decade building slick tools for accounts payable, the money going out the door. And it has minted multiple billion-dollar companies. Yet the exact mirror image, the money coming in, has been treated as a back-office chore nobody wanted to touch. It struck me as almost backwards. AR is the lifeblood of a business. It is literally how you get paid, and somehow it became the neglected side of the ledger. If Caitlin is right that the two are the same size, a lot of value has been sitting there unclaimed, and Fazeshift is well-positioned to unlock it. Anyway, that’s it for today’s show. If you enjoy these episodes, please go ahead and subscribe, tell a friend or leave a review. And thanks so much for listening.