$440 How AI is Fixing Late Payments: Fabio Fleitas on Automating Cash Collections

[00:00:00]

Mehmet: Hello and welcome back to a new episode of the CTO Show with Mehmet today. I'm very pleased. Joining me Fabio Fleitas. Fabio is the co-founder and CTO of Tesorio. I don't like to mention much [00:01:00] about my guests because as you know by now, I keep it to my guests, introduce themselves, tell us a little bit more about their journey, you know, and you know, their background and what they're currently up to.

Mehmet: So Fabi, without further ado, I will leave the mic with you. Just teaser for, you know, the audience, we're going to talk about everything finance for companies. And, you know, I will leave Fabio to, to discuss this with us. So Fabio, the floor is yours.

Fabio: Awesome. So, yeah. Um, as I mentioned, I am the co founder and CTO of Tesorio.

Fabio: Tesorio, for those who do not know, we are a connected financial operations platform that basically helps companies turn their revenue. Into cash, right? And that's fueling their own growth. And so how we do that is we are helping them with very specific tactical financial operations and workflows, right?

Fabio: Um, specifically, mostly focus on things that are helping with their cash inflows. Right. If you think about cash flows as cash coming in and cash going out, we really help around cash getting in. [00:02:00] Um, a lot of our workflows are around that. Things like collections, cash applications, payment processing and those types of things.

Fabio: Basically helping those finance departments, helping them with those workflows and helping streamline all of that in order to give them a better sense of their own cash flows.

Mehmet: Great. And thank you again, Fabio, for being here with me today. Uh, transitional question, legacy question. I know, but I'm curious about.

Mehmet: The story of every company, you know. Why it gets started. So I know like it's a vision. Sometime, uh, sometime it comes out off your own pain. So for you, um, you know, when you decide with your co founders to start the company, what was the motive? Like what is the pain that you have seen in the market that probably, you know, no one else was solving or maybe they're not solving it in the right way.

Mehmet: And then you said, okay guys, let's sit down and Build this. So I like to capture [00:03:00] these moments.

Fabio: Yeah. Yeah. I mean, a little bit of background. That's that's hopefully helpful context. Um, first, I started this when I was in university. So I and that's when that's actually where I met my co founder. So 10 years ago, I worked on this.

Fabio: So I was actually 20 years old. Um, when I first started to Soria, I am now 31. Um, so I've been working on this, uh, for quite a while. And, um, going into the university, um, I knew I, I was, you know, I was, I was pretty ambitious, uh, back then. And I knew I wanted to build a company. And I went to the university of Pennsylvania, which happens to have one of the best business schools and one of the best technology schools in the United States.

Fabio: Um, I went the tech route, but I was very interested in finding a partner to work with and try to build a company. And so it happened to be my co founder, Carlos. He was an MBA at the at that time in the Wharton Business School at the University of Pennsylvania. And, um, and so we met. And the funny, the funny background there is [00:04:00] that, um, you know, as a 20 year old back then, I didn't know much about it.

Fabio: the finance teams, the finance departments, and what they did, right? I actually didn't, uh, I, I actually, my, my, my partner, he, um, Um, he had had a decade in finance experience by then, um, and, uh, and really knew the space, but, you know, I had basic things like I didn't, I didn't know how business operated with invoice and stuff.

Fabio: I, I, I didn't know all those things, but, um, one of the things that, you know, we started working together is understanding kind of what's going on with these businesses and trying to build something together. Um, Fast forward a little bit. Uh, I want to talk about something, you know, we, we, we entered the, the, we entered Y Combinator, uh, in 2015.

Fabio: Uh, Y Combinator, for those who don't, do not know, is a startup accelerator out in Silicon Valley. Um, and we were fortunate enough to join that program in 2015. Um, raised a seed round afterwards and everything, which was, which was amazing. Um, but then actually as, happens a lot with startups. We ended up pivoting.

Fabio: Um, a [00:05:00] little bit over a year after Y. C. We ended up pivoting. Although we were still building finance enterprise products and features, we realized the initial one didn't work out. And so really, to kind of answer your question, we had to talk about the pivot, right? Because the pivot is when now, now what we're building and what we're working on is kind of came from that pivot.

Fabio: And, you know, a lot of people kind of asked me from time to time, like, Okay, like how do you do this? How do you think about this idea? How did you kind of work on this? Right? Um, and the thing about business software and business and selling to businesses is that I think it's easy in a logical way. It's not easy like Actually doing it, but using a logical way.

Fabio: Businesses have a problem and as, and if it's enough of a problem, a big enough problem, they'll pay you to solve it. Right? With software, something like that. And so what, what did we end up doing when we pivoted? We ended up literally having dinners with like finance executives in the Bay Area. We started kind of hang, like talking with them, understanding and like we just ask them, Hey, [00:06:00] what are your biggest problems?

Fabio: Right? Uh, if you're familiar with the book, the mom test. Um, which is a fantastic book. Uh, it kind of tells you like how to kind of interview, how to learn, how to understand your problems. So effectively we did kind of that approach really early on when we did that pivot. We're just like, we kind of went a little bit open minded.

Fabio: What were some of those biggest problems? We understood, you know, we had a sense of it, but we wanted to really understand from finance executives and leaders on that. And based on those conversations and everything, two things came out. One, we heard a lot, uh, that from them that the whole process of like, The collections process, receiving money, invoices, following up, applications, like all this stuff was just a nightmare.

Fabio: It was so manual, so brute force. Um, and then tied to that is obviously just cash flows, right? Because getting that money is the hardest part of cash flow. It's the most volatile. It's the most unpredictable. And so we're like, okay, let's go with that. Let's, let's do that. That seems like a big enough problem.

Fabio: Seems like no one's really solving that very well back then. And so we went [00:07:00] ahead and tackled that problem.

Mehmet: Fantastic. I love to hear these stories and, uh, Yeah, so pivoting is part of the journey. Um, you know, I, I mentor sometimes to be entrepreneurs and they get shocked. Ah, how come? And yeah, I guess start to give them examples from very well known companies.

Mehmet: I said, yeah, because you need to follow, you know, where your customer Ghana first. You know, use your product and pay for your product. So this is, this is very much part of the, of what every startup at some stage might have to do. Now, Fabio, one thing, which may be people who assume that, what, what are these guys talking about?

Mehmet: I'm trying to simplify it. And I know both of us, we come from technology, technology background, and I was shocked the first time I learned. So my assumption was, so if I'm a company. I sent you as another company, I sent an [00:08:00] invoice. Oh, happy life. You know, money is the bank. But it's not the case. So if we can highlight, you know, about the dynamics, of course, from high level, we don't need to go and give them a, a, like one or one course in finance regarding this.

Mehmet: But so for people, and by the way, for people who are planning to have a company, or maybe you want to start, you know, a company. So you need to understand, even if you are from technical background, you need to understand this. So Fabio, tell us a little bit about, you know, Okay. This account receivable account payable concept.

Mehmet: So we have more context.

Fabio: Right. Yeah, yeah. No, and that makes sense. So basically, you know, let's say, as you said, right, you're, you're a business and you just sold a product, a service, a good to another business, right? The typical kind of general process that you'll do there is, you know, you'll go ahead and sign an agreement, a subscription agreement, a master's agreement, whatever, that will say like, all right, I'm going to, you know, every month you're going to [00:09:00] build me this much, every year you're going to build me this much, right?

Fabio: Low amounts of money usually go on a credit card something but usually when you're you know Let's say you're sending a ten thousand dollar invoice a thirty thousand dollar and where's a hundred thousand dollar You're not going to get that on a credit card obviously because it's too expensive So usually what what happens, uh, especially in the u.

Fabio: s Other countries a little bit different but definitely in the u. s and a lot of kind of english speaking countries they'll put together basically a pdf an invoice pdf that says what what is uh, What is due? And when is it due? And what was it for? They'll send that, usually in an email, to the business on the other side.

Fabio: Someone will take that and then process it. And then, you know, hopefully, if everything goes well, pay you. later down the road, but there's a couple of problems along the way. One is the actual communications aspect, right? Obviously for one customer, one time you can do that. You can follow up. You can be like, Hey, just reminder it's coming due in, you know, in five days or in two days and follow up reminders when you have.[00:10:00]

Fabio: thousands of customers. And usually finance teams are just a couple of people, literally single digit people that I always find that shocking. Um, especially very large companies. I'm not talking about small companies. I'm talking about like public public companies. They have single digit, um, folks managing accounts receivable.

Fabio: They have to deal with hundreds or thousands or more, tens of thousands in some cases of customers and invoices following up, reminding them to get paid, dealing with issues, disputes, escalations, problems like that. And so our tool helps automate those workflows right before us. It was manual. It would you would manually we knew some of our customers just manually send these emails, manually follow up with customers and just deal with the whole the whole process.

Fabio: Now with the story of these, these folks can send those emails, automate the follow ups when the payment comes in, they can automate the reconciliation process in order to post that back into their accounting system and deal with with that whole process. What was historically a fully manual process, you know with [00:11:00] emails communications manual looking at bank statements Tessorio helps kind of automate that whole workflow

Mehmet: That's fantastic And just you know to your point because you mentioned about the u.

Mehmet: s and the markets that you operate in It's a bigger problem You know, in, in emerging markets. So for example, in the Middle East, let's say, because payments are like much slower and people sometime requires, like instead the standard in the U S I know it's like net 30 usually. So here we talk about, and when I say net, like, I mean, when it's the invoice due, so after 30 days, usually.

Mehmet: So sometime we hear we have 60, 90, 180 days. So, and. For the folks who also want to understand how this affects your business so you can go And listen to the episode. I had a couple of weeks ago with maliki holiday who you know Helps actually small medium businesses to get funds to avoid, you know being And I think fabulous.

Mehmet: One of the things also the solution [00:12:00] does is to avoid issues in cash flow. So how much cash they have in in the bank currently. Right. So let's talk about the financial aspect, like, I mean, who would be your ultimate personnels? Put it this way, who would benefit? When deploying your solution, Fabio.

Fabio: Yeah. So the types of customers that we work with, um, tend to be kind of mid to large, right?

Fabio: So we are, we haven't been targeting the very small kind of mom and pop type shops. Um, usually when you see this, when you have a dedicated finance team, when you have dedicated people managing accounts receivable, you know, you have, you have folks kind of in the finance space. And usually. That happens when the company is a bit more mature, right?

Fabio: They don't have to be public companies or anything like that, but they are a growing company, you know, in the startup world. And so, basically, the main value, as I mentioned before, that we help with is all this [00:13:00] manual and time consuming processes, right? All, all these financial processes and workflows that I mentioned, invoices, reconciliation, managing kind of your cash inflows, dealing with your bank accounts, all that has been historically manual before you using a tool like the Soria.

Fabio: So we can help automate that, um, for, for folks. And that becomes more important as you're growing, as you're scaling, because, um, you know, usually finance teams, again, I found it ironic. They're the holders of the budget. Right. They hold the budget in the company, but they themselves often do not have much budget, especially for people and personnel.

Fabio: And so, um, with a tool like this Oreo, they can grow, they can scale without having to hire additional, additional, more people. We've had some customers have been with us for years and I've told us they haven't had to hire anybody because, uh, they've been able to leverage disorders and just help that, that those people kind of grow and scale.

Fabio: And so, um, mostly scaling companies, growing companies. Our ideal customer profile, um, but [00:14:00] yeah all over all over the the world english speaking countries. We're pretty much every single one of them.

Mehmet: Perfect You know, I always also tell people and which is I can see direct benefit for Folks who use your solution fabio.

Mehmet: So first of all you Them streamline. Let's call it the collection process That's a

Fabio: big part of it. Yeah.

Mehmet: Yeah, so which is time and always I tell people time is money The second thing you give predictability on the cash flow I'm gonna ask you about that in a moment and also which I you know I'm a guy who come from you know, kind of a risk management it risk management or technology risk management So anything related to risk it attracts me.

Mehmet: So also, you know you help them avoid kind of risk mitigation of, you know, getting these, these collections lost, right? So, because it's, it's money now, which is fantastic. Now, again, it's not like I'm playing the devil advocate or something like [00:15:00] this. So do you get some time people asking, Hey, I get, you know, my ERP in place.

Mehmet: I get my, without mentioning brands, I get. You know, this accounting tool in place. So why do I need another tool? So this is the first question. The second one, how easy is the integration? Is it like something hard to do? Or is it something like seamless? Of course, I imagine using APIs or some kind of automation you would be able to do that.

Mehmet: So tell us more about that aspect.

Fabio: Let me, uh, one thing I'd love to kind of highlight to start off with, uh, to talk about is like, What is the benefit, right? Because obviously we could talk about efficiency and stuff, but like everyone will tell you that, right? Um, so I wanted to share a couple of stats that I think are, are, are very relevant.

Fabio: First, I'm going to talk about a general kind of stat across kind of our customer base, uh, right? As well as then a specific stat from actually a very specific customer of ours that I think, uh, has been, uh, has been with us. for [00:16:00] very, very long time and, and has been very successful with us. So, you know, across our customer base, there's a couple of things that we help with very, very specifically for finance teams, right?

Fabio: Uh, day sales outstanding. That's one of the metrics that, um, companies, uh, finance teams are looking at, which is basically how long does it take to get money in the door. Uh, and so what we see, uh, uh, on average is 33 days production, uh, for our, for our customers, which is, which is. Which is a very big metric.

Fabio: Ultimately, what does that mean? Reducing the amount of time that you receive cash unlocks your cash from your balance sheet, right? So all that cash that you're sitting there is outstanding invoices that are delayed. Our delinquent is impacting our cash flow. And so we've been able to unlock 200 million dollars from the boundaries of our customers.

Fabio: The first 12 months of their use because of that, right? So forget the productivity increases, which we're seeing around three X. We're seeing companies be able [00:17:00] to use the cash, use their money, um, by leveraging a toilet story by helping them hit their targets. We've had, I remember there was one company that implemented it.

Fabio: Um, public company and they were already planning to tell, uh, their board and their investors that they were not going to hit their targets, their collections targets. Um, and, uh, within the same quarter that they happened to implement Tesorio, they literally went from about, they were, they were going to tell their board that they were not going to hit their target to exceeding their targets within that short period of time.

Fabio: They literally told us like, guys, like I want you, I want you to hear this because this, this was, this is an amazing story. And so they told us that that happened. Um, and so one of the customers I want to mention specifically, you know, there's been with us for a very, very long time. One of our earliest customers actually is a Viva.

Fabio: Uh, they're a cloud based software company in the global. Life sciences industry. And so with, with a partnership we've had with them over the last several years, they've, they've been able to do a couple of things help with their [00:18:00] finances. One, they've been able to reduce their bad debt write offs by 75%.

Fabio: Right? So what does that mean? That means those, those invoices that they were going to have to just write off because they're like, We can't, we're not be able to collect on it. They were able to reduce that by 75 percent using Tesorio. Also, if you look at how old invoices are, you know, if they're 30 days overdue, 20 days overdue, 60 days overdue, et cetera, right?

Fabio: Overdue. They've been able to get their 90 days overdue, 50 percent reduced, right? Um, and what does that mean? That means that they can get cash in the door, right? They can get cash in the door that, um, that helps them grow their business and reinvest in their business versus having to go out and get financing, going out and using other means to get cash.

Fabio: Um, so operating that way.

Mehmet: Wow. I love to hear these stories, you know, because I'm big believer always. And I say, you know, technology should be a driver for business benefits. And this is exactly the use [00:19:00] cases that I, I wanted to hear from you, Fabio. Um, it's. Kind of mix of technical slash business question for you.

Mehmet: Now how, and I'm going to ask a quick question based on it. I will decide if it will be relevant to ask the second. So how much you have to do with integrating with the banking? Solutions of the customer.

Fabio: Yeah, it's a great question. Um, for certain parts of our product is necessary. For example, specifically the cash application side of things.

Fabio: So cash application just for those who do not know, is the process of basically, you know, it's cash application reconciliation. Both are exactly what we're doing. Um, but basically it's taking, you know, when a customer pays you often, it's just an ACH, a wire money just shows up in the bank, right? You have to go in and like, look at your bank statements, and then look at those, and then see which invoices they were closed out for.

Fabio: Sometimes your customer will send you an email, [00:20:00] sometimes they won't, you know, and so you have to kind of manually apply the cash that came in to the invoices. Uh, within your system. So in order to do that, for example, we do need connections to banks, right? Um, so we do have connectivity into bank data. Um, for that, as well as kind of helping understand another part of product is just being able to understand your bank balances, your cash flows.

Fabio: Um, and so using bank data in order to power that as well.

Mehmet: Yeah, great. So my other question is, you know, because you've been doing this for now almost 10 years. Yeah. Again, maybe 11 now. So Things in in fintech as we call it today. It's changed a lot since since then and Now we have the new generation of neo banks and you know, like these kinds of the bank that is trying to Digitize everything, you know, so have you seen this something that you know pushed what you're trying to [00:21:00] do forward.

Mehmet: Was it like something which put more hurdle because you needed to wait customers, you know, to, to adapt with what these banks are trying to do? Like even traditional banks, we know now they're trying to become the neo banks, right? So from, from that perspective, if, and we focus on the technology aspect of it, Fabio, like if you compare When you started with today, what would be the biggest differences you saw?

Mehmet: Especially I'm talking about banks because money has to come to the bank. But of course, to, to also share, you know, some breakthroughs that happen within, um, you know, the ecosystem.

Fabio: Yeah. Yeah. Um, so on the banking side, there have been some. Changes that are beneficial for us. I'll give a caveat to that in a second.

Fabio: Um, obviously, you know, if you look at the in Europe with PSD two and stuff that's been forcing, uh, European banks to open up their data, uh, you know, mostly via APIs for consumers, um, that did [00:22:00] force also, um, banks in the U. S. to do, to do, to do the similar thing, right? And so what you're seeing is, um, within the U.

Fabio: S. and other countries in Europe as well, um, access to banking data becoming easier. The caveat to that, you know, and, and actually what one, you know, As, as, as folks may know, like, you know, there's big players out there like plaid, uh, that make it, you know, their one stop shop API for all bank connectivity.

Fabio: And you know, even when they started, they had to do a bunch of screen scraping and stuff because APIs for these banks didn't exist. This is something that banks have been slowly rolling out. That said, there is still one challenge, um, that we still face, which is that most of these APIs are for the small business or consumer.

Fabio: Thanks. Right? So, you know, you as an individual can connect your, you know, account, uh, into into plot or whatever tool you're using, but, uh, you [00:23:00] know, a company like chase or whatever, they have the consumer bank accounts, they have the small business bank account, then they have the enterprise accounts, right?

Fabio: And they are different, uh, they are separated. And so, although we are seeing more and more banks, uh, at least traditional banks, uh, exposing their APIs, they often still are lagging for enterprise ones. And then the other thing still, uh, you know, with these neobanks and stuff, um, what we've seen is obviously quite heavy, mostly on the consumer side.

Fabio: There are obviously business ones, if you look at like Brex and Mercury and stuff that are, you know, targeting, uh, a lot, obviously startups, um, with their banking, banking services. But, uh, Again, for better or worse, the types of customers we deal with, often they have to kind of graduate from those banks that we see a lot of traditional banking, um, in our, um, ecosystem and our customer base.

Fabio: And, uh, there are data challenges there. [00:24:00] Accessing bank data from. The enterprise traditional banks is still challenging to date. It's getting better, but it's still it's still far behind consumer banks.

Mehmet: Cool. Super cool. It's good that you mentioned data here because the next part I want to discuss with you, Fabio, taking into consideration the regulations and everything that happens around.

Mehmet: So building machine learning models and AI models within your space. What are the challenges and what are, like, the use cases that you have, you know, implemented in Tesorio that, you know, you can tell us about?

Fabio: Yeah, I know. I mean, you can talk a lot about AI. Um, so even before generative, just quick, quick little segue, little step back, even before generative AI, we have been leveraging machine learning for helping predict when invoices will be paid, scoring customers based on payment behaviors, we had had kind of machine learning built into our product from very early on.

Fabio: That said, Generative AI, uh, [00:25:00] did change the game, um, for us in, in, I think, very meaningful ways. I think there are a lot of and bold claims, uh, from some companies around AI that I think are still not really, you know, I think, I think are still early for some of the claims that they're doing. Um, I'm excited to see the, like, AI is, you know, continuously, um, improving and getting cheaper.

Fabio: Um, and I'm excited about the progress there. But the good thing is we found actual good use cases for generative AI within Tesorio. But to kind of answer one of your questions, uh, in particular, you know, We're dealing with finance. We're dealing with financial data. You know, we're working with, um, traditionally the more conservative part of, uh, uh, a department in a company, right?

Fabio: The finance department, they, they tend to be a little bit more conservative, a little bit more risk averse, et cetera, right? And so we've had to take that into consideration as we're [00:26:00] building out our generative AI capabilities. We are building, um, a lot of new features on new products, but we've taken a measured approach to that one.

Fabio: That's a very important one is we give customers control over what AI features they want to enable, disable by default, everything is disabled, right? So customers can enable the things that they feel comfortable with the features that they would like and sharing the data there. Obviously, we are sock to compliance, we have all our security, we work with a lot of enterprises.

Fabio: So we're very confident in all those practices. But Obviously, there's still some companies that are still a little bit averse, uh, around adoption of AI and want to kind of wait it out and see how things go. And that makes sense. We're, we're happy to oblige there. So customers have control over what they, what features in AI they use.

Fabio: And then, um, Oh yeah, go ahead. No, no, go ahead. Uh, and then another important part, uh, which is always kind of. You know, been a little bit of the scary part of, uh, of AI is people thinking about like, Oh, it's gonna steal my data. It's gonna use my data, [00:27:00] right? It's always been a little bit of a scary topic. I think it's been getting better and better since the, you know, since the launch of chat GPT, basically, um, and enterprises are getting more and more comfortable with it.

Fabio: And so, um, One of the things that we do around that is making sure that people understand what data is being shared. We do share the minimal amount of data, and ultimately, we do not train on data. That's the most important part, right? A lot of companies are worried about using their data to train models, to do things that, you know, may cause problems for their businesses.

Fabio: And so we, at the end of the day, are AI capabilities are all to serve a customer, the customer that pays us, that uses us, that uses our product. We're not in the business of building foundational bottles or anything like that. So data security and AI have always been an important part from our, uh, uh, perspective.

Fabio: And we continue to kind of, uh, invest in there.

Mehmet: Cool. Um, you AI and generative AI for Tesorio. [00:28:00] I believe this is something, some of it, because it would be related to You know external stakeholders, maybe your customers, but what i'm curious about like being a cto also fabulous Um, there's a debate about you know, utilizing it for coding And you know, you know, you know They are saying now you can use it as your junior software engineer.

Mehmet: So what's your point of view on this?

Fabio: I have, uh, actually a decently strong point of view on this. Um, so I guess, first of all, I've been spearheading within our company, the use of AI, like I've been pushing it really hard. Um, I think AI is a tool that will make everyone's job easier and everyone more productive and more effective.

Fabio: Um, and so I have been pushing it very heavily within Tesorio in order for us to adopt it and I've been doing a [00:29:00] couple things to do that. One, um, we, uh, everyone in Tesorio can use Chad, GPT, or Claude. Right. So we pay for team accounts for that. And so everyone can join, um, uh, to leverage GPT or cloud.

Fabio: There's slight differences between them that, uh, that folks may have a preference for. So we support both. We also, for our engineers do provide AI code editors. Um, Uh, such as cursor, get a copilot, um, and so we offer a couple options and if folks have one that they prefer as well, happy to do that. Um, I really do think that this, the, the, the trend that we're on right now saying you do not want to use, uh, an AI assistant, whether it's a chat GPT or code editor or whatever in order to do your job would be in my mind, the equivalent of saying like, I don't want to use the internet.

Fabio: You know, like, back in the day, like, it's, I think it's, it's a crazy statement, it's a tool, right, that can help you, and yes, just like any other tool, you can't just go blindly, just trust [00:30:00] whatever code it generates, whatever it says, like, you do have to actually still do work, uh, in order to verify all that, but it is a very strong tool, and so one of the things that I've been doing with, with my team is, um, We have every two weeks a full kind of product, dev, design, engineering, uh, all hands, uh, effectively.

Fabio: And during there we do, we go over announcements, updates, things like that. One of the things I introduced is, uh, everyone in the company has to, has to fill out a form. How have you used AI in the last two weeks, basically. So everyone, uh, says that. And then I look through those use cases and I asked them, I was like, can you show this in the, in the all hands, like basically, and so people get to showcase that.

Fabio: And that's been. So cool to see how people are leveraging it. And so then people get inspired. People are sharing kind of templates. People are sharing prompts, uh, in order to do that. And we've been seeing things from like people auto generating, you know, PR descriptions and titles really quickly just by providing the diff.

Fabio: On a github [00:31:00] commit and stuff like that so it can generate you like a title and description there, uh, and so we've been seeing all these use cases. I think I'm such a strong proponent for leveraging these tools. I feel like anyone who is opposed to it will just fall behind. Honestly, it's like it's a It's a thing that yes, it's not perfect.

Fabio: And I think people think too high. Like they think that, you know, you install curse or get a copilot suddenly, it's going to code for you the whole feature. It's like, No, it's not. But like understanding how to leverage this tool, I think will become a requirement for engineering going forward, to be honest.

Mehmet: Absolutely. I like, you know, uh, when you mentioned like, it's like someone who refuses to use the internet, um, because literally, you know, like when I use almost all, all of them. So chat GPT from open AI, uh, cloud. Gemini. I'm not a coder, but we recently installed

Fabio: Gemini on our G suite [00:32:00] as

Mehmet: well. And, you know, out of curiosity, although like I come from technical background, I, I don't do coding, but just out of curiosity, I tried a bunch of these, uh, tools like cursor, like there's now a bunch of tools that came out recently that they do fantastic job.

Mehmet: And, you know, I sit down, I say, okay. This is really makes the life much easier because I put myself in in the place of a developer or like software engineer and say, wow, like it's writing for me, or it's at least doing for me repetitive tasks. And I think this is one of the main, uh, main things that I kept telling people about from technical and business perspective.

Mehmet: Um, you know, there's this, uh, side of the house that they keep nagging. Yeah. The, yeah. The content is like pretty much similar to each other. I'm talking about copywriting and all this stuff. And I say, guys, I think you're looking at a very narrow angle where you are just looking [00:33:00] at, uh, these tools for your marketing slash sales.

Mehmet: You know, if first don't get me wrong, if you write a good prompt, can I give you a good answer sometimes? Yeah, but, but it can do much more. And now with the agentic AI, like agents and operators, like the one that, uh, uh, open AI, they have released that we're recording at the beginning of Feb. So just two days ago or yesterday, you know, open AI, there is this, uh, deep reasoning, you know, agent also as well.

Mehmet: So let me ask you Fabio this. So with, with this rapid change as a CTO, as someone who's putting. Of course, you have the business vision as as a company, but for you, it's like driving the technology. And yeah, you're pushing people to use AI and so on. But from, you know, bigger vision perspective, do you feel sometimes it's going too much fast?

Mehmet: Or You know, you would say it has [00:34:00] always been in tech like this, so nothing to worry about.

Fabio: I, I don't think it's going too fast. Um, I think, I know there's, there's considerations of risk on, on, uh, making sure that we're doing it safely and stuff. And I, and I do think companies like OpenAI, Anthropic, et cetera, are taking good steps there.

Fabio: So, um, for what it's worth, I, look, I remember. Um, several years ago, I would be part of some, I would, I would go to AI conferences and go to AI talks and stuff. And I remember I always walked away. You know, they would always talk about the potential for AI. They would talk about, like, oh, it's gonna change our lives.

Fabio: It's gonna, it's gonna change the way that we build things. It's gonna change the way that we, we operate as a society, etc. All these crazy bold claims. And I was just like, guys, when is this gonna happen? Because, like, you know, we're basically just [00:35:00] doing machine learning and statistical models right now, right?

Fabio: Like, outside of some really complex Uh, things that very large businesses, we're not doing anything like wild here, like, in my opinion, right? Like, yes, big data side, what is producing is like, you know, what is my Uber arriving? When is my delivery arriving? Like, it wasn't anything crazy. I feel like, like, in terms of what is it chad GPT changed that, right?

Fabio: We finally saw a glimpse of like, oh, wow, like, yeah. This is real, like we can really do stuff that like, you know, what we, what we always kind of dreamed about what AI could be is finally more real than it's ever been before. It was always this like fantastical, almost concept several years ago. And now it's been progressing at such a pace that you're like, wow, look at all these things, right?

Fabio: Operator, as you mentioned. Um, Is, is, is, looks fantastic. And, and again, [00:36:00] um, from, I think one of the things you were kind of talking about there is how do I think about this for Tessori? How do I think about it from a technologist? I think everything that's going on is amazing. One, Not only are models getting better all the time, right?

Fabio: Like we're seeing advancements, multiple advancements every year. Yes, we haven't seen that big leap like three to four yet. Then, you know, but I says maybe we won't see a five. I don't know, but we're still seeing significant advances and significant changes in the stack, right? Um, and the way that these tools are being used.

Fabio: That is amazing. We're and I'll be excited to talk to you about how we're already leveraging some of these newest technologies for our own products and what we're doing there. But it's also getting cheaper, right? Like every time it's just getting cheaper and cheaper in order to do these things. And so that's allowing Developers is allowing people to love to have kind of like, like turning a I almost into a commodity, [00:37:00] which I think is beneficial for society.

Mehmet: Absolutely. Uh, it seems fabulous. Like, you know, we are on the same page because for me, um, I remember a couple of months after ChatGPT, you know, went viral and everyone started to use it. You know, you remember there were, uh, this, you know, these people who were trying to slow down things. Yeah, I think they

Fabio: got banned in multiple countries in Europe.

Mehmet: Yeah, and I said, guys, like, what are you trying to do? So it's just out there. And like any other technology, of course, it has you know, some disadvantages in a sense, like related to regulations and, you know, some people can use it in a bad way and all this stuff. But I said, yeah, but the internet was no different.

Mehmet: You know, a lot of other technologies, even non technology stuff, it's non different also as well. So everything you can use it like in [00:38:00] a mixed way, but yeah, to your point. And I agree with you that we're not going, of course it's, it's fast, but I mean, I'm not worried about it being fast because I think we, We waited a lot because I remind people always like the AI thing concept machine learning Started, you know a century ago, right people trying to figure out Sorry how to You know, you tell a machine something and the machine just figure it out and do it.

Mehmet: So we're not 100 percent there yet, but we're very close to this moment, especially what we are seeing with the agents. That's why I'm very excited about the agents. I'm excited about the operator. I'm excited about, you know, everything that's going on. I don't care. About, you know, the, I don't know how to call it, but you know, like some, you know what, like, for example, deep sea game and everyone was saying, uh, open AI is done.

Mehmet: So guys, they're not, I think I, people are not seeing the big picture [00:39:00] is. Already, you know, we have unleashed, you know, the, I call it the genie out of the bottle and all what we are seeing now is just like, you know, the after effect on this. So I don't want to let you go because I, I'm afraid that I will forget to ask you because you mentioned about some of the use cases that you said you want to mention.

Mehmet: So tell us more about that.

Fabio: Yeah. So again, I, I'm a huge, as already mentioned, I'm a huge proponent of, of, of AI and generative AI in general. Um, and I think You know, once we started to take a look at, like, the capabilities of OpenAI's Chad GPT and the GPT models and stuff early on, we actually built, um, our first, uh, generative AI feature a couple weeks after Chad GPT was announced, like, we were like, wow, like, you know, I think, um, we, we, we, we started seeing some potentials within, within our product, but let me kind of tell you a little bit about what I think generative AI means, uh, for us to soar as a business.

Fabio: I think [00:40:00] it unlocked capabilities that were previously were just not there. possible. Really? Like you're just way too difficult. Um, and we've ended up building both copilots and agents to help with financial operations. So let me kind of break that down. I think one of the one of the things to think about when you think about financial data is that there's obviously the cleaned up data that comes from accounting systems, ERPs, CRMs and stuff.

Fabio: But a lot of financial data exists in unstructured formats that we think about as primary source data, right? There's a lot of data in emails. There's data in documents, right? PDFs, signed order forms, even the invoice. PDFs and stuff. We're seeing also because a lot of financial communication happens still with emails.

Fabio: It's not, it's not some crazy digital thing still, right? And a lot of it's just emails back and forth between people, basically. And, um, and so with our, um, regenerative AI, we've been able to do a couple things. One, we've been able to now [00:41:00] tap into this data. So, for example, um, one of the things that we currently have in our product is, um, there's a lot of systems that generate, um, notifications around payment statuses, invoice statuses.

Fabio: So, when you submit an invoice, it'll tell you it was approved, it was rejected, um, the money's on the way, here's when it's coming, et cetera. Like, just data coming from these systems on the status of these. As you can imagine, if you're sending hundreds or thousands of these per week, you will miss these notifications.

Fabio: But with one of our generative AI, we can actually read these emails and structure that data. So basically that's actually powering one of our workflows, which is basically monitoring your inbox and taking out information from it that's relevant for the collections process. Things like when is payment When our payment is going to be where something disputed was something even approved for payment, and then we could even, you know, based on that approval, we can actually update what we think is the date that it will be paid, helping them understand when their cash was going to come in, understanding their cash flows.

Fabio: So [00:42:00] we're able to now read emails, we're able to read, uh, Uh, this data in order to structure it and then, uh, operate workflows within the Sorio. We're also able to read documents, PDFs, sales orders to generate financial data, things that usually literally required human data entry in order to do this.

Fabio: Right. And now AI can take this document, extract the relevant data to create the financial document automatically versus having a human have to read it. And like traditional OCR and technologies like that were Honestly, that like you have to really fine tune these OCR models for specific use cases versus here, uh, with generate, you could just change your prompt a little bit to change the description of, of what that field is and boom, you can extract it much more effectively.

Fabio: We've been seeing much success on that now. Yeah, go ahead. No, it's okay. What? One other thing. Uh, Okay. that I'm really excited about that's [00:43:00] coming soon. Um, we actually built, uh, an, uh, an operator agent also within Tesoria. Um, so basically what this operator agent does is, um, one of the things obviously mentioned, uh, around the collections process is all the email communication reminders and stuff, but there's one big piece that we've, I've always missed.

Fabio: And it was because it was just not really possible. to do this, um, in any meaningful way, um, which was, uh, oftentimes there would be customers that had a portal that you have to log into in order to fill out fields and submit an invoice, right? So instead of emailing them, you had to go manually log in, drop down, click, fill out the form, attach the PDF, make sure the tax was correctly added.

Fabio: So basically they're like outsourcing. The accounting work to the person submitting the invoice, and that trend is becoming more and more common. Like we're seeing that grow. We're seeing more adoption of these, uh, accounts, received accounts, payable portal supplier portals [00:44:00] that our customers have to go in.

Fabio: and submit their invoices to. Now there was no way for us to deal with that before because it required literally logging into these portals. There was no APIs, at least there are no APIs for these or anything like that. But we have been able to successfully build an eight, uh, an operator agent that can literally log in, find the invoice, do the purchase order, submit it all on your behalf.

Fabio: Um, and so that is actually literally in beta with several customers right now that's doing it for them. And it is. Exactly what you saw, like with open as operator, but for this specific use case, which is submitting invoices into these portals, right? So, um. Um, and it's, and the model and the work that we've been doing and the prompts that we're doing are specifically tuned to do that work, basically making sure we fill out all the fields, making sure we're double checking it in order to submit these invoices.

Fabio: And that's, that's, that's really cutting edge, like the way that we're approaching that problem.

Mehmet: Absolutely fantastic. You know, and this is why I tell [00:45:00] people, um, keep, you should expect this. Of course, it's, it's fascinating Fabio to hear this from you because I can imagine how much time has been wasted before on these repetitive tasks.

Mehmet: And let's be also honest with ourself, we do mistakes, right? So filling the form, like how many times, like you try to, to fill. You know, similar force or maybe even like such a similar form for for, you know, getting payments and you go back and forth over email, sometimes over the phone to fixing a thing that it should get in seconds.

Mehmet: I would say not even minutes. So this is fascinating to hear. And it's nice to see that you have adopted such, you know, cutting edge technology in your product. And I hope like people get inspired and understand because this is not only about the technology itself, Although, like I [00:46:00] said, fascinating technology.

Mehmet: I keep repeating fascinating because it is, um, but it's because you make people lives easy. You make business. you know, more streamlined from finance perspective. So this is really fantastic and thank you for sharing this with us. Fabio. Um, as we're coming to an end, I still have like two questions to ask you, Fabio.

Mehmet: So for people who are in tech and they want to be in the fintech game, right? Uh, anything related to finance. Um, what do you advise them to have? under their belt so they can understand because I can, you know, imagine that working specifically with anything related to financials is not like smart to working with any other thing.

Mehmet: So what usually you advise, uh, people who are still in the early stages of their career to get [00:47:00] done in order to become like, you know, technology leaders.

Fabio: So this is, uh, just to confirm, uh, make sure understand this is mostly for Uh, for folks in the technology role that are interested in getting into into fintech, right?

Fabio: Um, no, it's a good question. I mean, I would honestly say that there isn't anything specific that you would need to do there outside of your general kind of technology role. Right? Um, so I, like, for example, the engineers that we hire, the product managers we hire, and stuff like that, uh, often don't need to have a finance background.

Fabio: It is beneficial, obviously, understanding the financial operations, the financial workflows, and stuff. It definitely does help. But, um, more so than anything, right, especially if you're mostly in the software engineering side and everything. [00:48:00] A lot of the, uh, I think I would say the B2B. Enterprise sass experience is the most relevant part, right?

Fabio: Obviously, there is a there is a difference between consumer application building consumer applications versus business applications. There is a difference there. But as long as you have the business. And you're building financial software like we are for businesses, then a lot of kind of the enterprise business software business sass, um, knowledge that you'll have will transfer over.

Mehmet: Great. And. What advice you would give to entrepreneurs in general who want to be in this space now, and especially with all the AI things that we talked about. So I like to call it, you know, words of wisdom for these folks who are either You know, started their journey or about to [00:49:00] start their journey in entrepreneurship.

Fabio: Yeah. In this

Mehmet: domain.

Fabio: Yeah. Yeah. I mean, generally speaking in the entrepreneur, like entrepreneurship, irrespective of this domain specifically, um, I would say there's a lot of AI tools coming out that are making your jobs easier, right? There's a lot of tools that's helping you create sales copy. There's a lot of tools that are helping you build a marketing website faster.

Fabio: There's tools that literally build. Product help you build product UI. There's obviously code editors to help you code. I guess I would say irrespective of which field you go to, you need to understand the AI tools ecosystem in order for you to be effective because at the end of the day, as an entrepreneur, the most valuable thing we have is time.

Fabio: Right? And and unfortunately, you never have enough time ever. You never feel like you do. So at least make sure that you're leveraging tools that help you move faster, you know, help you build a prototype [00:50:00] quicker, help you write code quicker. I actually met a buddy in Barcelona a couple months back, who was had only around one year of experience.

Fabio: And with him having, like, for example, using cursor, it was amazing. See how fast fast. He was able to build something with so little experience. Um, he had built a couple of like prototype apps and stuff that he was trying to test out. He was a, he's an entrepreneur as well. And, um, I, I literally was blown away.

Fabio: I was like, wow, like, you know, um, with so little experience, you can actually get a lot done nowadays. So that's, that's important for generally for entrepreneurs in terms of this space, you know, um, I think the most important thing is kind of what I talked about. Uh, when we pivoted, right? When you're dealing with businesses, when you're dealing irrespective of its finance, payments, whatever, always make sure you're solving a real, real problem, right?

Fabio: It's not one that you think it's a problem. It's not one that you might, you know, you [00:51:00] hypothetically could be a problem. literally go talk to users. I think again, one of the best books that helps you understand how to interview users is the mom test, right? And then it shows you kind of frameworks for how to do that.

Fabio: Fun fact, actually, one of the things I still use the mom test today. Um, it's really easy to understand the mom test. You know, it's kind of like, it's kind of like, you know, it's really easy when someone explains something to you, but like then generating it is really hard. So one of the things I ended up doing a while back was having a chat GPT probably like you're an expert at the principles of the bomb test and you'd help creating questions for this, you know, and so I use this for products that we were building, not a whole company, but like a product, right?

Fabio: Um, and so that's a little tip there. It's like, you can actually use Jeju Q Cloud to help you generate the types of questions that would be relevant to make sure you understand if your business idea has any grounds.

Mehmet: Super cool. Yeah. And, uh, I got a. Make sure I keep [00:52:00] the link for the book so people can go and grab it.

Mehmet: Um Fabio, you shared a lot with us today and you know, I learned a lot myself where people can Stay in touch or keep in touch and find more about you and about Tesorio

Fabio: Um, so for Tesorio, um, you can obviously check out our website, www. tesorio. com. We actually recently revamped it, um, quite a bit. There's full product tours.

Fabio: You can actually see all our products. There's ways for you to kind of interact with our products. So, um, obviously if you're interested in learning more about what we do and, and the types of things that we offer. Uh, our website was recently refreshed and it looks great and has a lot of those things. Um, for me, um, you can mostly check out kind of links and stuff on my website.

Fabio: It's just FabioFiletest. com. Pretty simple. Um, and then there you'll see my LinkedIn and other things like that, that are relevant for me, but from there you should be able to find everything you

Mehmet: need. Great. And again, thank you very much Fabio for being here with me today. I really enjoyed the conversation [00:53:00] and I'm sure a lot of people will benefit whether businesses who needs that or, uh, fellow entrepreneurs, uh, in general and people who are interested in.

Mehmet: Um fintech specifically, so thank you very much for being here with me today And this is for the audience. This is how usually I add my episodes. So if you just discovered our podcast by luck Thank you for passing by. I hope you enjoyed it if you did so, please You know subscribe share it with your friends and colleagues who are available on all podcasting platforms and we are available On youtube also as well.

Mehmet: And if you are one of the people Who keep coming back. Thank you very much. And I'm saying this since, you know, approximately three, four episodes because one month ago, uh, the podcast as CTO show with Mehmet. We were humbled to become top 40 most listened business podcast in Dubai. So thank you for the support.

Mehmet: And also we've done something we never done before. So we became also in the top 200 chart, actually top hundred [00:54:00] chart in six countries simultaneously. So something that never happened in the past two years since I started the show. So thank you very much. This couldn't happen without your support and your encouragement.

Mehmet: And as I say, always. Stay tuned for a new episode very soon. Thank you. Bye.

Bye.