Carla Canino. Pod

Monica Millares: [00:00:00] They are living $443 billion on the table, not because of fraud, but because we are not lending to the right customers and we're declining good customers

Carla Canino: The bank flagged me as high risk, and it required me to physically go authenticate myself in the branch. But at the time, I was unfortunately in a wheelchair. I have a disability and I could not physically go authenticate myself, which meant their KYC system and their fraud system.

Again, blocked my account and I could not access my account for months.

To a certain degree, bots will be our customers.

We're thinking just from an operational cost perspective, a small institution say of 2 million customers will have an operational waste of between four to 5 million annually just on operational mismanagement, which is crazy.

Vast areas of the population not having [00:01:00] access to the same access to financial tools, but it literally is costing the industry billions. And if you don't have a quantification, if you don't have measurable output, it just gets brushed under the table. So we need to speak the same language. We need to speak the language that matters, and that language is money.

resilience, it's not that I decided to be a solo founder, it is that I'm the last man standing all the time.

Nobody's gonna give you permission. No one, like truly no one's gonna say to you, yeah you should be a solo founder.

Right? It's like the kind of thing where you do it because you have to do it.

But the truth is, you just need to go. And the recognition comes and the respect and visibility and acknowledgement comes, once you create value.

And sometimes you have to toot your own horn.

Monica Millares: In today's episode, we are going to go through a question that many lenders ask [00:02:00] that it's like, how can we lend to more people without losing money? Because it's like at the top of our minds all the time, right? Like, how can I expand but without relaxing my roles and keeping keeping the rates under control?

So Cara, welcome to the show.

Carla Canino: Thank you so much for having me. So excited to be here. Happy New Year. Wishing you also health, prosperity and tons of positive impact.

Monica Millares: Thank you. Yes. That's a good one. Tons of positive impact for everyone. Yeah, exactly. And talking about impact.

Carla Canino: Yes.

Monica Millares: Basically there's a stat that you've shared before that it's basically lenders across the world.

They are living $443 billion on the table, not because of fraud, but because we are not lending to. The right customers and we're declining good customers as such. [00:03:00] That is a lot. It's 65% of applications are declined, 443 billion worldwide. It's a very, very big topic to discuss, but before we go into that,

how should we as managers basically. Tackle this question of there's this opportunity, we want to lend more people to lend to more people, but if we do, basically the portfolio becomes more risky and we lose money. So what's your take on how do we reduce false declines to expand the pie?

Carla Canino: It's a big question and it's an excellent it's an excellent position to be in.

Why? 2025 and onwards the implementation of artificial intelligence in various use cases. Financial technology has, of course, exponentially expanded, right? Like the investment of artificial intelligence is now [00:04:00] the, the broadest investment in in the financial stack. What does this mean? It means that the legacy decisions and the legacy assumptions and the legacy classifications that we have made historically in how we viewed customers are not only going to continue but actually going to be exacerbated and exponentially in many cases.

If you were talking about. Biases and the decline of good customers is going to exponentially worsen. So when we're talking about a 65% of customers is good to understand that it's 65% of declines of the current customer declines are actually worthwhile customers. What does this mean? This means that we have an issue with misclassification of risk and we have an issue also of understanding who is a good customer.

That also means that the very models that we use to classify and to evaluate customers are inaccurate. And if we create or if we invest in a stack that [00:05:00] continues to make the same decisions in the same way, we will continue to turn away good customers. It's important to understand also this core question, which is, who is our ideal customer, and how many of those are there?

So when we take a step back and we look at customer segments and the reality, the human reality of our customers, most models are designed, or most models are actually evaluating for customers that don't really exist. If we look at, I personally work with understanding the prevalence of what we would call underrepresented populations.

So

Monica Millares: mm-hmm.

Carla Canino: About 23% of the population worldwide has some sort of disability. Yeah, about 20% of the population worldwide might be elderly. About also 19 to 22% of the population worldwide might be neurodiverse. Obviously about 49 to 51% of the population, depending on where you are, might be women and so on, [00:06:00] and we have about a 15% representation depending on where you are of immigrants.

So all of these, when you aggregate these customer segments and, and, and more, you actually understand that your current customers and your future customers are. Vulnerable populations are underrepresented in these models, which means, again, the models that we have created are simply mis misclassifying risk.

Monica Millares: So is this what you mean with the 443 billion stat, or is that different?

Carla Canino: Yes. So the 443 billion is the representation of the missed opportunity in both. Lending and the total addressable market, right? But the problem is actually broader. So when we are designing product experiences and we are making, again, product assumptions and risk assumptions about customers, this doesn't just affect lending, which is the funny thing.

Other areas that Jason use cases such as KYC, underwriting, [00:07:00] customer service, the very flow product flow that we use to onboard customers. And that affects not only the potential revenue, but also operational cost. So things like if you have a customer that already has a debit card or a credit card with a low limit, and they don't have the ability to get appropriate customer service because they might have a disability or because they might not speak English as a first language, that means that you're.

Most in fact, financial technology firms are under serving these customers and creating friction that costs in terms of operational costs, in terms of escalations, in terms of manual escalations, in terms of, of course, things like refunds, and of course, in customer trust, which reduces the share of wallet of future, of, of current and future customer.

Monica Millares: Interesting. Because then basically, if I were to summarize this, you are saying we are doing product design for the, just like we have happy path and on Happy path, we have product [00:08:00] design for the quote unquote happy customer. Yes. Eg. The prototype that fits all the tick boxes.

Carla Canino: Yes.

Monica Millares: But we are not designing for all these subcategories of customers.

Hence we're living money on the table.

Carla Canino: Absolutely. We are not only leaving money on the table, it's costing us more already in operational experience. So again. The path that we have chosen, or the design that we have is not only misinforming our data banks, right, and it's mis flagging customers, but it's also already costing in manual escalations in unhappy customers, sometimes canceling their accounts or chargebacks or refunds.

And of course, in the case of lending. It's mis flagging people or creating additional friction even in the, the onboarding or the application process that mis captures information that again, creates risk or, sorry creates flags around. Customer behavior that the model [00:09:00] itself just simply does not understand.

I think it's just for context, it's, it's good to explore The reason why I ended up creating kindly. My background is, of course, in financial technology. I'm, I come from the payment side and the fraud prevention side, but when I first moved to the Netherlands, I opened a bank account with a local bank.

And funny enough at the time my, my income signal was at least twice what the. Local average is so it wasn't an income problem. The bank flagged me as high risk, and it required me to physically go authenticate myself in the branch. But at the time, I was unfortunately in a wheelchair. I have a disability and I could not physically go authenticate myself, which meant their KYC system and their fraud system.

Again, blocked my account and I could not access my account for months. And this was of course in the beginning of COVID, so I couldn't really go anywhere. I wasn't going to expose myself. So [00:10:00] these were simply adjustments that could have been both from a customer service perspective, from from a accessibility perspective, and of course from an understanding of customer behavior that could have been managed differently.

I, of course closed my account eventually. I was so frustrated with that bank that I took my money elsewhere. And so this bank will very likely never see me, not only open another account, but of course, you know, capture any kind of other share of wallet that I could have potentially had, including lending.

So I think this is just one of the many different experiences that what we would call underrepresented populations might have.

Monica Millares: Yeah. And then you work. A lot within the lending space, like your customers work within the lending space. So who aren't these customers that are being declined?

Carla Canino: That's a great question.

So we recently created our first benchmarking study by analyzing customer service interactions with different financial technology [00:11:00] firms. So it spans beyond lending, and we analyzed a few different customers segments. So typically when we talk about, or we think about. Misrepresentation or, or the, the very controversial term bias.

We tend to think you know, either about gender or about race, but actually models misrepresent a variety of different customer segments. So again, we talk about disabled people. They might have a visual impairment or they might have a. A mobility issue, they might be also elderly. And so elderly people interact with technology very differently.

They might have low visibility as well. We have neurodiverse people like me that have 300 different tabs and forget that they have a session open and might forget to come back to it, end up getting kicked out of the session. But you also have a variety of other. Again, we go back to gender, we go back to race, we go back to immigrants that might not have English as a first language and might not understand the the requirements or the [00:12:00] particular flow of a new country.

Right? And so all of these segments will interact in the financial system a little bit differently than locals, for example, right? And they might have also different access to identification or might have a. Thinner customer file or might simply not understand the right steps or provide different data or documentation that required.

And a lot of times, even if you're, if we're talking about things like. The onboarding process and the biometric authentication required for a lot of these things. Some customers might have either actual disfigurements or again, their skin tone might not be as well recognized by a biometric authentication or in my case, and some people, or a lot of people have little fingers or renal syndrome or a worn, so that means that the.

Decline rate for authentication and for biometric authentication is much higher for some of these underrepresented groups. And so all of these things as an aggregate create additional friction that [00:13:00] end up not only frustrating the customer and creating operational costs, but truly end up declining perfectly good customers that we should have taken into account in both the customer journey and the data and the, the data story and the of the risk itself.

Monica Millares: Yeah. That's very interesting because as you, as you were describing these customers and, and specifically the example about the hands, I was like, yes, I never think about that. Like, right. Embarrassing, but, but that's the reality, right? It's like, I don't, but then I was, then my BC product mind came in and I'm like, no, because even if we did, it's not the top priority.

Hence we couldn't allocate, basically resource to fix that.

Carla Canino: Yeah.

Monica Millares: Because of a big backtalk that we have. So how do we, I guess like if I'm a neobank, how do we measure the impact of not, not serving these customers? Well,

Carla Canino: I think the ROI question is truly what we focus [00:14:00] on and what we care about at kindly.

So our thesis is that by reducing some of these disparities and some of these points of friction, and by shedding light, we can not only, of course, help. Customers be better represented, but also create a higher profitability for, for fintechs and also obviously be able to create a clear path of higher revenue.

There's a couple of different cost aspects that are important to, to acknowledge here. As we were discussing before. The operational cost aspect. So it might be very much escalations or might be require or requests for accessibility that quite frankly a lot of financial technology firms are falling short in that aspect.

And that of course means that we are. Eh, unfortunately under serving a lot of these segments, but it also means that we need to actively analyze and actively actively evaluate how the [00:15:00] models are performing in production. So it's not, not only, it's not enough to analyze how your models are performing in house.

Or as you're developing them, I think there's plenty of responsible AI solutions that provide a certain degree of help in that, but the reality is that you want to actually measure how a lot of these models are performing to real customers. The other aspect of this is that as a bank is developing their solutions and their risk models.

Most banks are also depending on foundational models for their, their, their LLM base, right? So they really cannot control not only the guardrails, but the performance or whatever security or data injections might happen on that side. And so it's extremely important to be able to look at how the. Models are performing on the other end, right on, on the model in perfor, in in production, and that's exactly what we are focusing on at Kindly.

So we evaluate [00:16:00] how these models behave on essentially. Synthetic vulnerable personas, and we are able to quantify how this is actually affecting from both an operational perspective and from a total addressable market gap. In addition to that, we are able to provide a certain degree of recommendations so that the financial technology firm can.

Retrain their models and improve their representation towards these population segments. And so again, that not only creates a clear visibility on the cost aspect and the potential uplift that they could gain, but also it creates a path towards improvement and towards representation, which gives them hopefully also a competitive advantage.

Monica Millares: Yeah. So before we go deeper, can you explain to us what does kindly do? Because like you call it the trust and compliance layer for AI in finance. What does that mean? [00:17:00]

Carla Canino: Yeah, that's a good question. So again, there's, there's a lot of solutions right now that can, at least on a high end, evaluate or help build you know, a certain level of.

Representation for customers. But what we do is we take it a step further. We check how the, again, how models are performing in production. And we have specialized models that can better represent and measure how, how the financial technology firm is actually providing full access and representation to a lot of these groups that are underrepresented.

And we provide cost visibility. We co, we also provide a full compliance measurement, so. We care about not only how, you know, the financial technology firm is, is performing from a profitability perspective, but also how they're performing from a compliance perspective as the AI Act is coming. And also the European Accessibility Act, at least here on the European side, that's, that's something that was enacted recently.

It's important for them to have [00:18:00] full visibility on how their to. Not only ISO 2 7 0 0 1, which is more around transparency, but also ISO 2 4 0 2 7, which is very much around how are these companies performing from a bias reduction perspective. And that's very much kind of like the flag that we analyze in addition to things like accessibility and fairness across many different customer segments as per the definition by the European Commission.

Monica Millares: Okay, so if we were to summarize and if I am a new bank and I run ai, AI powered credit decisioning, how do you help me?

Carla Canino: Yeah, we provide a full suite of analysis from, again, from a bias perspective, from a compliance perspective, and from a cost perspective as well as the full suite of recommendations.

So you can retrain your models and reduce the friction as well. So hopefully that will better represent customers and create higher profitability for the company. [00:19:00] And

Monica Millares: Okay. And then you also say, this is like the skeptical me.

Carla Canino: Yes.

Monica Millares: Speaking. You talk about, you have a core promise that it's zero integration.

Carla Canino: Yes.

Monica Millares: I'm like, no, come on.

Carla Canino: That's a,

Monica Millares: how does this work?

Carla Canino: That's a, it's important to say that we are, first of all, an early stage company and our first. Use case evaluation. Use case is actually looking at how companies are performing from a customer service experience. 'cause that's the first point of interaction, right?

From there, we will be expanding into other use cases that are relevant across the industry and that are part of the evaluation ultimately, not only for lending, but for different core needs of the, or core uses for, for, for banking. Such as you know. Money movement. And so what we are able to do is do what's called adversarial de biasing through our own models without the need to integrate.

So I can't get into the specifics, [00:20:00] but we understand, like we come from the, the FinTech world, we understand that PII is something very delicate. We don't want to expose real customer data, but we also want to. Create visibility. So we are able to do this by essentially leveraging our bespoke models against the the customer service databases and the customer service experiences.

And, and give at least a first view in the future. We are hoping to expand the use cases, as I mentioned before, into KYC and other use cases that are a priority in the industry. And I don't know if we can maintain that promise of no integration, but we do understand that low lift is something that is required.

So fair, fair.

Monica Millares: Okay. Cool. Good. I'm like, nah,

Carla Canino: no, but it's not necessary. It's a funny thing, right? We don't. We don't really wanna expose, or, you know, frankly, I would not, I don't want to touch real customer information if it's not necessary. And it's truly not necessary in order for us to be able to give you a good view on [00:21:00] whether you are at least in, in the first layer, in the first line of interaction, whether you like.

Whether the bank is, is truly treating customers fairly and and understanding the needs of what we would call, we used to call outliers on our under represented customers, but we actually realize that they're actually the majority of the customer base.

Monica Millares: Okay, so then it's easy to integrate.

Carla Canino: Yes, it's

Monica Millares: easy to implement.

So then let's go a little bit deeper, coming back to the use case of lending. How do you exactly help us basically approve more customers without being more risky?

Carla Canino: Yeah, so when we, when we're thinking about lending, first of all, this is a, this is a complex answer because the approval models vary depending on country, right?

And we also have to understand that when we're thinking about biases, a lot of times people think about prejudice. But in this case, we're talking about friction. We're [00:22:00] talking about misclassification, and again, misclassification of risk, and we're understanding, we're also talking about bias, omission bias, meaning we simply don't have representative data.

We don't have the full visibility of true customer experience. And so what we are doing is helping contextualize the needs and the customer journeys and the real life experience of a lot of these underrepresented segments precisely so you can better represent them and have models that are next generation, have models that are actually accurately representing the experience of a lot more customers while also.

Understanding that, and this is important. There is no perfect scoring meaning. I think it's unrealistic to think that we will solve. You know, and, and have a, a, a zero false declines. I think that we can have improvements by better representing some of these journeys, as I mentioned, right? Like such as my own journey.

And we do that by [00:23:00] looking at the different areas. That are involved in the lending process. So lending is not only the last scoring on customer credit worthiness, it involves, as we said before, right, like customer journey. It involves customer service, it involves onboarding, it involves KYC, and so in biometric authentication.

So there's a lot of pieces in here that can be better contextualized, and that's exactly what we're aiming to do. We think that in order to solve this problem is not a, how do you call it, a a zero to a hundred solution. It's very much about incremental improvements that better represent and better contextualize real user experience.

And it's quite complex because again, it very much depends on not only how a particular country might manage or might require or might view credit worthiness, but very much about each financial institution and how their own tech stack performs also, [00:24:00] versus. The external tech stack that they're depending on.

So right now, one of the things that we are able to view that most companies can't do or can't view, is the performance across the industry. Right? So we launched the kindly. Cost of trust benchmark, precisely because we want to be able to have a, a true temp temperature of how not only the different financial institutions are performing, but also how are they performing vis-a-vis the, the LLMs or the, the, you know, the external tech stack that they're utilizing.

So we can provide a better ROI in a better view, in a more true to customer view on who are they depending on. If you're using OpenAI versus Gemini versus mistrial or you know, whatever data vendors you're using, ultimately it's very difficult for you to compare or for you to view how they're performing unless you have something like kindly truly viewing or giving you a view on how you're benchmarking [00:25:00] against the industry.

Monica Millares: Interesting. Yeah.

Carla Canino: Is that helpful?

Monica Millares: Yeah. It's like, I hadn't thought about that, that it's like Yeah. The, the benchmarking against the industry. Yeah.

Carla Canino: I think ROI is one of those questions that are right now very important in the industry, right? Like a lot of, as we said, like a lot of companies are investing in in ai, but what is the ROI and how do you measure for that?

So that's one of the core questions that we have in mind is not only how do we help the population have a fair access to, to financial tools, but also if the financial world is essentially moving to be as AI driven as possible, essentially in the future. To a certain degree, bots will be our customers.

So how do you know what is the effectiveness? What is the reliability and how, and what is the drift and what is the true performance in production of your tech stack vis-a-vis the customers that you're trying to serve?

Monica Millares: I love that. So in the future, I, I love that you [00:26:00] said that it's like, in the future, bots are going to be the customers.

Carla Canino: Yep. Yes.

Monica Millares: So. That takes us back to compliance. You spoke about the European AI Act as such. What's the relevance here?

Carla Canino: That's a, yeah, it's a tricky one, right? Because we've seen, let's take a step back and, and this is a tricky one because of geopolitics and because of, you know, sometimes what can be described as the over reliance of certain industries in US tech and vice versa and, and not vice versa.

Sorry about that. An overreliance of certain industries in US Tech. The AI Act was very much developed to both protect the sovereignty and European values and European priorities towards European customers, but it also is providing. A certain degree of transparency around how US technology is being leveraged and used and how it's performing exactly in those terms, right?

Like how is it affecting European [00:27:00] customers? And the same thing can be said around the world. Like if the, the vast majority of customers of US technology are act actually outside of the us, then it's A-A-A-A-A-A high highly demanded, gimme one second. This one is hard to explain.

So if a high degree of the clients of US technology outside of the us, then it begs the question, how is US technology serving the citizens of other countries, especially in highly regulated industries such as financial, te. And so and, and, and I think we all are aware of the Brussels effect, right?

Like whenever you have regulatory introduction like ISO 200 22 a few years ago that tends to influence how other countries adopt regulation and how they view risk. And so the AI Act, I think it's a [00:28:00] really good example of, regulation that is at least, and, and of course, imperfectly trying to create transparency for its citizens while also trying to create an environment where European technology is rebirthing, right?

Like it's, it has a fighting chance. And so I think that kindly provides a lot of this transparency and provides you know, a very unique value, which is. The specific specialization on ISO 2 4 0 2 7, which is again for bias reduction, but it also, I think it's providing a very clear. Transparency around economic access for populations that are typically not served.

And it, it truly pushes the industry to quantify the value of these segments and to quantify or to see these, these segments as again, competitive advantage. So I believe that we have underestimated. The, the value of real people. Right? And [00:29:00] by definition, we have created models that are exclusionary and kindly is able to create not only transparency around that, but a, a true cost of opportunity measurement that doesn't really exist anywhere else.

Okay.

Monica Millares: So now it's like that's the money question.

Carla Canino: Yes.

Monica Millares: I'm like, okay, tell us more. Tell us more.

Carla Canino: Yeah. Yeah.

Monica Millares: So yeah, what are the, what are the economics? Is it a SA model, like revenue share? Like

Carla Canino: how do we charge it?

Monica Millares: How does it work? Yeah.

Carla Canino: Lemme backtrack a little bit. First, let me say why this matters. So, in the preliminary study that we created, and again, we are a very, very early stage.

So this was done across 10 technology firms. Identified that in every 10 interactions, eight had had extremely high friction. That's really bad, right? So again, out of eight I'm sorry, out of 10 customer service interactions, eight, [00:30:00] had extremely high friction, including the control. So if we, if we actually take a look at how that affects a given financial.

Technology firm. We're thinking just from an operational cost perspective, a small institution, say I'm thinking about US decisis, but an institution say of 2 million customers will have an operational waste of between four to 5 million annually just on operational mismanagement, which is crazy. Right.

And so

Monica Millares: $5

Carla Canino: million? Yes. Just on operational mismatch, and this is the first time that we're doing these tests. We are not looking at KYC, we're not looking at again, lending yet. We're not, we're we're, we're truly looking at what this is credibly costing. Year on year to a small institution. So if you extrapolate that this is a, a huge cost, so why does this matter?

First we can, [00:31:00] we have to pull back a little bit to the compliance side. So the AI Act requires a third party, an independent third party to perform these these analysis or these evaluations. So that's one of the reasons why we're doing what we're doing. The other one is. We don't know what the drift is going to be from a bias perspective in production to financial technology firms, which means we have to run randomized tests essentially monthly and for a period of a year at minimum, in order for us to not, not only be able to accurately quantify, but actually help a financial institution improve.

So we do have, yeah, an audit fee. And for certain types of financial technology providers. We do have a revenue share model, which we'll be introducing later on this year. But for again, a typical financial technology firm, when we're, when we're, when we're talking about the high risk models as per the AI act, the main quote unquote cost or the main investment would be [00:32:00] continuous audits.

In addition to that, we again provide. Their recommendations so they can retrain their models and thus reduce a lot of this friction and reduce a lot of these operational waste and hopefully, again, gain competitive advantage.

Monica Millares: So then to recap that, let's say if I'm a neobank and I'm assessing kindly, what are the leading indicators that I could have to track to then say, Hey, this is having impact on the p and l?

Carla Canino: Yeah,

I

Monica Millares: think working with kindly.

Carla Canino: Yeah, yeah, yeah. Kindly has a really unique position because we're able to provide visibility on essentially all, all quadrants of the p and l, right? So as I mentioned before, we have direct visibility at first, even without integration, we have direct visibility into the cost aspect of things.

We can extrapolate from that. Also, we can extrapolate from that also how the revenue uplift can be affected in terms of. Both customer experience and also at the total addressable market. [00:33:00] We have an influence in terms of improving bank's representation of their actual customer base. And we also have a full visibility and full can provide a a, a full compliance and which obviously affects their risk quadrant performance.

So we provide full scoring. On these aspects of the AI Act as well as the EAA, as well as some aspects of GDPR that require, once again, fair treatment.

Monica Millares: Interesting. So then if I'm assessing kindly, there's always a question of, oh, can we build this ourselves? Oh yeah. As a product person.

Carla Canino: Yeah.

Monica Millares: It's like, oh, why should we, should we clean house or should we go for external?

Yeah. What's your logic behind it?

Carla Canino: Yeah, I think it's necessary for co, for, for companies to have in-house responsible AI tools. I think it's like, as they're building, that's something that we as kindly actually want them to do. We want them to improve. [00:34:00] However, a couple of caveats, so we come, we, we are essentially, we sit on top of the tech stack, right?

So we once again evaluate. In production performance. The other aspect of this is that we have aggregate data and we have benchmarking that gives us visibility that obviously, you know, even if you do build your, your own tech stack, you wouldn't have. And that's part of the value that we provide. And we have, of course, specialized models and specialized technology that does this.

That being said, what again, we go back to. How are you going to know what's a true ROI if you only depend on one particular tech stack, right? You don't really have visibility of how you're performing vis-a-vis other companies or how, what is the true ROI of your technology if you are reviewing all of these tools and you are also dependent on your very own evaluators.

So the, the funny thing here is that the majority of the, the evaluators that exist right now are developed by the companies that we need to evaluate. [00:35:00] So if you think about Fair ln, if you think about ia a BM 360, all of these are evaluators that are widely used in the industry as companies develop.

They're, that that's basically what a lot of these companies use which ironically enough need to be evaluated afterwards because they simply, they're skinny, right? Like they do not truly. Have the, both the, the data, the data with, and the true representation that is needed to measure this. But also it's essentially impossible for them to be able to evaluate both, you know, their own data stack and have a true contextualization of what's happening across the industry.

Monica Millares: Of course. Yeah. So it's kind of like you cannot be the patient on the.

Carla Canino: You got it. Exactly. I'm, I'm thinking about the Spanish. Like, you, you cannot be Judge Verdugo. I don't know what verdu means, how you say that, but Yes, it's exactly that. It's exactly that. And that's exactly why we want to remain independent, right?

Like the idea here is for us to be able to provide a true, [00:36:00] independent and unbiased view and, and be able to provide recommendations based on what we're observing in the market and based on the company's specific gaps. Oh, the, the proper context. So I'm, I'm doing my best.

Monica Millares: Yes, exactly. Yeah. Which then that takes me to you as a founder.

Carla Canino: Oh my God.

Monica Millares: You know, like you had a thriving career in financial services. Yes. And then after 15 years, you were like, you know what? I'm going to go and become a FinTech founder. What was your thinking and decisioning framework? Can you guide us for that process?

Carla Canino: Yeah, so while I had a career or I still have a career in financial services, technically in parallel to that, I worked a lot in, in financial inclusion.

So I worked with the advisor Federal Reserve for a little while. I also collaborated with the W three C, and I had also essentially a serial founder failed. But, you know, it's always a [00:37:00] journey, right? Like, and both the companies that I tried to, to, to take on were related to accessibility. So it's always been part of what I do, right?

It's it's always been very much my desire and it's one of those things that I think is in your veins. When you see a problem, this is also why I want, you know, I've been both in the business development world and in the product world because you're able to, like, you know, if you identify a problem, you want to solve it.

And that's, I think the core of product is very much understanding friction, having empathy for customers, and trying to find a solution for that. And so I think it's just like a compulsion that I have and maybe a slight level of masochism. For, for kindly. You know, I think my personal experience, as I was mentioning when opening an account, but also every single conversation that I have around this topic, right?

Like with people that are immigrants, like there's so many of my friends that are immigrants that have moved and, and in different socioeconomic brackets have had this same [00:38:00] issue, right? People that even also work in financial industries have had issues opening accounts, right? I of course know a lot of other people with different kinds of disabilities and they don't have representation.

They are constantly underestimated and they're exposed a lot of times to more fraud because they have dependency on third parties to be able to access financial tools. You know, obviously everybody has parents, so every time that I talk to, you know, anybody and, and you know, we're discussing how AI is now.

Being adopted across the board. Everybody's concerned about how their parents are going to interact with, with technology. Everybody's concerned about them having to help their parents or their parents being exposed to again, you know fraud. And so I think it's very much, for me, it's very much a mission.

Monica Millares: Hmm.

Carla Canino: It's is, I think it's clearly an opportunity and to me is the biggest risk area, both from a security perspective and from an industry gap. And I am [00:39:00] seconded by, you know, other people in the, in the industry, including people that lead compliance and lead risk and data analysis. But it's very much a mission, right?

Like. Somebody has to do it and I, I feel like I'm in the privileged position of both having the experience and the perspective, a very unique perspective that I have gained both through being an immigrant, but also by being disabled, by being a woman. I mean, I'm obviously Venezuelan that, you know, it would almost be.

I think it would be criminal if I did it, if I didn't try to do something. I really do think I'm compelled to do it and I'm in a very unique position to, to be able to educate and be able to create value. Right. The other aspect of it is one of the most difficult. Paradigm shift and, and, and challenges to the industry is this concept of what are biases and are biases a cost?

Do we really have biases in our models? And these segments of the population are really like small. And again, you know, I think reeducating the industry and [00:40:00] provide. Concise and measurable quantification of the impact, both from a cost and opportunity perspective is absolutely necessary, is vital, because not only the underestimation of of these customer segments is going to lead to.

Vast areas of the population not having access to, to the same access to financial tools, but it literally is costing the industry billions. And if you don't have a quantification, if you don't have measurable output, it just gets brushed under the table. So we need to speak the same language. We need to speak the language that matters, and that language is money.

Monica Millares: It's, yeah. And like the industry financial services as more, as much purpose driven that we want FinTech to be. It's at the end of the day, it's, it's we are businesses, money

Carla Canino: industry. Yeah. It's a business. But as business, we need to put on the business hat and say, well, no, actually these, these segments of the population are [00:41:00] huge.

I mean, they're not, they're not getting smaller. Right. Like in many, in many developed countries, the population is aging. So if you're not developing tools that take into account the needs of aging populations, then you're going to create even more friction and you're going to create even less trust, and that, again, has a quantifiable impact on the valve.

A financial services firm, and I mean, this is also a segmentation problem, right? It's very much a customer experience, a product, and a data impact problem, because if we are not able to truly understand who our real customer segments are, then we can't create products that are, that are actually serving the right.

Needs of the customers. And so how are we gonna be profitable? Right? And, and what, what risk models are we creating? How are they accurate if they're not actually understanding the customers? So it's, it's, I think it's a core need. If I, I cannot think of a, of a problem in the short to medium to long term [00:42:00] that matters more than what kindly is addressing.

Monica Millares: I love, I love, not only, I love how passionate you are about it, you know that it's like. Yes, and we're going to go and solve it. Which you are so passionate about it that you decided to become a solo founder.

Carla Canino: Yes. Yes.

Monica Millares: What was your, yeah. What made you be a solo founder? Most people are like, Hey, fundraising, I need you to three people.

You were like, no, forget about that. I'll just go solo.

Carla Canino: I think resilience, it's not that I decided to be a solo founder, it is that I'm the last man standing all the time. Like, again, I've had a couple of, I have had a couple of previous startups, right? And, and by virtue of many things and this sounds a little bit like tooting my own horn, but I have a very particular resilience profile and again, a, a very particular set of a, a very particular perspective that is able to inform both the product and the direction of the company.

And [00:43:00] so I think that also I am incredibly mission driven. It's very hard to, to find other people that match that description. It's incredibly difficult. So it's not that I want it to be a solo founder, it's honestly that I, at least so far, I haven't found people that have the same level of commitment that would be able to put the company ahead of everything that understand what we're trying to solve and have a skin in the game to the level of that I do.

Right. To me, this has to be solved, and that's a result that not a lot of people have. And also, again, like. I think that that question actually stopped me and slowed me down for a long, long time because I did not think, and this happens I think to a lot of women, but to a lot of people, right? Like it's very scary to go and be a founder.

It is painful. You will have to solve a million things, and you need resilience. You need adaptability. You need, you know, to be incredibly mission driven and believe in what you're doing, and it's very scary to go by [00:44:00] yourself. And so. Precisely, that slowed me down so much until I tried to have a few people on board that would not pri prioritize the company or that had a different idea of like, you know, what biases meant.

And, and maybe they were a little bit too idealistic and not pragmatic enough. And so it's, I think it's very important to understand your own value and understand your own profile and how you can, you know, like truly. Value what you have and go with it. Nobody's gonna give you permission. No one, like truly no one's gonna say to you, yeah, you should be a solo founder.

Right? It's, it's like the kind of thing where you do it because you have to do it. And here we are and so far that, that essentially makes you what you need to be. Like. You just go and like approach the path and face whatever you need to face, and then you become the leader that you need to be.

Monica Millares: Amazing.

And in this journey, of course. Yeah, you have to speak with customers. You had to [00:45:00] make decisions When it comes to prioritization and product build. Product design, was there any moment that you were like, this was the thing that made me pivot into this way?

Carla Canino: Yeah, for sure. Quite a few times. I think the first time was so kindly started as something completely different.

I was actually analyzing a, leadership journeys and the leadership acumen for HR recruitment, right? And then the AI Act was being developed and then I realized, well, of course, like my background is in financial technology. I took a step back and, and saw again what we're discussing today, like how many different points in the customer experience in, in.

You know, not only for lending, but in general, have I been misclassified or have I cons, have I faced, or people that I know have faced challenges, how many different areas are going to be impacted by this wave of AI [00:46:00] and where can I help people the most? And I think that was the first pivot into financial services, right?

And, and understanding that. There was not just one use case, but there were a plethora of use cases that impacted not only one segment of the population, but everybody, and, and this was exactly at the same time that the AI Act was being developed. So I decided to pivot in that direction. And the second one was more recently, deciding where to start was very difficult, right?

Like there's so many different use cases in financial technology where you can analyze biases, but how can we quantify this if our, if our thesis is. There are biases. We need to quantify them in a way that matters to financial services and that doesn't open a can of worms and like, make them retreat.

Because the first conversations that we had were financial services firms saying, we don't have, we don't have biases. And then lo and behold, you know, six months later, a lot of them have been penalized because of misrepresentation, because of [00:47:00] active and in this case discrimination, right? So a little bit more, more grave.

We started looking at how can we do this in a way that we don't have to touch PII in a way that we can easily quantify the points of friction in a way that we can actually create value for not only the consumers, but also obviously the financial technology firm. And that's when we decided, let's, let's look at how the the customer jour journey goes and, and how.

Are these customer service bots, which is like essentially the core investment in AI that is being driven right now in addition to internal efficiency. How is this the first line of interaction actually working? That gave us a very direct view on exactly this, on like how are these models and how are these banks actually representing customers and actually treating customers?

And that opens the gate to a much broader conversation. So those were like the core moments of like Aha, that I've had so far in this journey.

Monica Millares: Amazing. Coming back to, you're a Soda [00:48:00] founder right now. There's this narrative everywhere online that AI companies, they are going to be like, you know, like Soda founder, soda founder, soda founder.

So for everyone listening, FinTech is working in startups that are like, I want to be one of those. What's your advice for people who are in the, I want to be a soda founder. I want to do something with ai. I'm not a, I'm not an.

Carla Canino: Mm.

Recently I was talking to a friend who has a huge network. He specializes in a very specific area of financial technology, and he has a vision for what he wants to do, but he didn't think that he had the, the engineering background necessary to create something. We always. We always have insecurities around our lack of skill or perhaps our lack of credibility like I did, right?[00:49:00]

If you have a growth mindset and if you have the ability to figure it out, you can do anything. If you have resilience and adaptability, you can do anything. So focus more on the how can I, and what did I learn and understand that there will be pain along the way. This is just a given. And try to convert that painting into lessons and try to convey that, convert that painting into strengths, and you will be able to do anything.

There are tools, and if there aren't, you can create them even if you don't know how. Being a founder is not the same as being a builder. You can hire builders, you can become a builder. You your core job is to figure it out. Your core responsibility is to figure it out, to create an amazing team around you, to be able to create the values and the culture that you need to be able to understand the problem, not the product.

[00:50:00] And actually I think it's an advantage, right? Like, 'cause a lot of times engineers create, like people have an engineering background. I have a product background, so I don't, I don't necessarily say that I'm non-technical because I do have, I mean, I've worked with engineers for the last 15 years. But I'm not, I'm not an, I mean I'm not a developer, but a lot of times people develop technical solutions and fall in love with the solution and not the problem, and they get stuck.

And the reality is that a successful company is not the tech stack. Right. It's the ability to best solve a problem and to find a wedge and to find, you know, a positioning that actually solves a problem with a customer. And the ability to have a customer that was, that is willing to pay for it, right? So I think that's that's one of the core pieces of advice and just like, jump in the water and swim.

Monica Millares: Amazing. Cool. So you jumped in the water.

Carla Canino: Wait. I'm a decent swimmer. A decent swimmer.

Monica Millares: So if you were to go back in time, [00:51:00] what's the decision that you could do differently?

Carla Canino: Oh man. There's so many. I think. I think I wasted a lot of time doubting myself and wanting people to be co-founders. Actually, I wasted a lot of time relying on other people to give value instead of me just going for it.

I've wasted a lot of. Time doubting myself and I mean, I think everybody goes through this, right? Like, like maybe investing in things that you didn't need to invest on and you could have like be been more frugal on one side and maybe invested on something else. Like these are just, everybody goes through this.

Even large corporations where you understand where your time and your money are best served. But I think those are the, the, the core areas where I would. Change. Yeah. And I think I wasted a lot of time also wanting external validation. When by definition I have experience that a lot of people don't have and they can't see [00:52:00] it.

And, and I think this goes for a lot of people, like in my particular like skill stack is not only in financial technology, but again, inclusive financial access. Like I understand the behavior and the pain points of a variety of segments of the population. And I've lived not only in Europe, but obviously I, you know, I'm American, I'm Venezuelan, I've also worked across Asia.

So I feel like I was sitting there waiting to be recognized, especially as a woman. Or getting defensive and getting angry that people didn't see me. But the truth is, you don't see people. You're like, you just need to go. And the recognition comes and the respect and visibility and acknowledgement comes once you create value.

And sometimes you have to toot your own horn.

Monica Millares: Yes. Two. Two. Amazing. It's been. Great. Speaking with like, you've got, you've got like a full year ahead. We wish kindly like the very, very,

Carla Canino: thank you so, so much. I'm so [00:53:00] excited. Thank you for the opportunity. I'm so, I'm always glad to talk about financial inclusion and this is, this has been a wonderful conversation, a wonderful way to start the year.