Jason Rosen, CEO of Prism Data, on how cash flow underwriting is transforming lending today

In this special episode recorded live at Money2020 in Las Vegas, we dive deep into what I think is the most transformative development in lending over the past decade: cash flow underwriting. I welcome back to the show Jason Rosen, CEO and co-founder of Prism Data (I last had Jason on the show back in 2019 when he was with Petal), to explore how real-time bank account data is revolutionizing credit assessment in ways that traditional credit scores simply can’t match.

Jason explains why the financial disruptions of 2020, from pandemic relief to widespread job changes, exposed critical gaps in conventional credit reporting, and how cash flow underwriting is filling that void. The conversation covers the breakthrough adoption of this technology by major traditional banks, the limitations of FICO scores in capturing a complete financial picture, and why cash flow underwriting represents a once-in-a-century shift in how creditworthiness is measured.

In this podcast you will learn:

• How Jason first got interested in cash flow underwriting.
• The origin story of Prism Data and how it was incubated inside Petal.
• How he describes Prism Data today.
• What goes into building their unique credit models.
• How their CashScore is created.
• What matters most in how they distill all the semi-structured data into a score.
• Why credit decisions are more intuitive when being made with cash flow underwriting.
• The lending categories where they are seeing the most rapid adoption.
• How smaller financial institutions can avoid being left behind.
• Jason’s view on why traditional credit models have become less predictive.
• How lenders are using cash flow underwriting in their application flow.
• How the friction in connecting a bank account is being reduced rapidly.
• The impact of the Chase-Plaid deal on cash flow underwriting.
• What sets Prism Data apart from the others in the cash flow underwriting space.

Read a transcription of our conversation below.

FINTECH ONE-ON-ONE PODCAST NO. 561: Jason Rosen

Jason Rosen:

One of the advantages of cashflow underwriting that we haven’t touched on yet is that it’s more or less real time. You’re able to see information from the last 24 hours of transactions. With credit reports and scores, there is a significant period of time before new information gets furnished and incorporated into those things. We really saw this in the financial crisis first and then again during COVID and the aftermath. At the end of 2020, you had millions of Americans that were out of work, that couldn’t go into their jobs. FICO scores hit an all-time high. Now, lenders then were smart enough to know that people hadn’t reached a new level of credit worthiness in that moment. So they knew that credit scores weren’t telling them the whole story. But the problem was, in the absence of some other reliable source of information, it left the industry flying blind and access to credit sharply declined.

Peter Renton:

This is the Fintech 101 podcast, the show for Fintech enthusiasts looking to better understand the leaders shaping Fintech and banking today. My name is Peter Renton and since 2013, I’ve been conducting in-depth interviews with Fintech founders and banking executives.

This episode is a little different as it was recorded live at Money 2020 in Las Vegas with Jason Rosen, the CEO and co-founder of Prism Data. Now I last had Jason on the show back in 2019 when he was with Pedal. So he explains how Prism was born from his experience with cashflow underwriting there. We also talk about how cashflow underwriting is experiencing a breakthrough moment in 2020, being adopted now by some of the largest traditional banks, as well as dozens of fintech companies. Jason explains why traditional credit scores are becoming less reliable, how bank account data can provide real-time insights that credit reports miss, and why he believes this represents a once-in-a-century shift in how credit decisions are made in America. Now, let’s get on with the show.

Okay, welcome to a special live episode of the FinTech 101 podcast. We are here at Money 2020 on the afternoon of day three and I’m here with Jason Rosen, the CEO and co-founder of Prism Data. Jason, so how’s the show been for you? You’ve been here since Sunday, right? And what’s it been like?

JR: I’ve been to a lot of Money2020s at this point, probably more than I, than I care to admit. I think this has been the busiest one.

PR: Wow. So you’ve been having lots of productive meetings, I take it.

JR: We are experiencing a perhaps once in a century shift in how credit scoring works, how credit worthiness is measured and determined in this country. And in the midst of that, there’s a lot of conversation to be had, a lot of productive activity and change in this space.

PR: Let’s get right into it and you’ve sort of teed it up nicely for us here. I cashflow underwriting is a hot topic right now. We’ve seen a lot of announcements. We’ve seen a lot of momentum, it seems, from lots of different players in the space. And you guys are really one of the pioneers in cashflow underwriting. I don’t know if anyone was doing it commercially before you guys. I know some banks were doing it internally from what I’ve heard in the past, but it feels like you were, you really in some ways invented the category. maybe you can take us back before we get right into it to why you decided to focus on cashflow underwriting and how, what was it like in the early days?

JR: Yeah. First, it’s amazing that cashflow underwriting has the momentum and the buzz around it that it does today. Certainly over my whole tenure of working on this, this is actually the minority experience. Most of the time, cashflow underwriting was an obscure side project that people weren’t paying very much attention to at all. So it has stepped into the spotlight more recently. We started working on cashflow underwriting circa 2015, 2016. So it’s been about 10 years since the very beginning.

Like a lot of entrepreneurial stories, we started with a personal pain point. One of my co-founders came to the States as an international student and he didn’t qualify for credit because he lacked a US credit history. His background was somewhat unique in that he was studying data science and machine learning, went on to get a PhD in machine learning, data science, and I saw opportunities to apply his academic research to the issues that he had experienced in his own personal financial life.

We began with a pretty broad exploration of how the credit system could be modernized to incorporate more information, new technologies and statistical techniques that would allow us to make better decisions in service of, you know, ultimately getting to the right outcome more often, right? Creating a system that works for more people and that gets to the right answer more frequently. We did a pretty broad exploration of the world of quote unquote alternative data. So all of the information that is not incorporated into a traditional credit report and credit score to evaluate which data sets could help to fill in the gaps that exist in the incumbent system. And that led us to look at a lot of different things. mean, there were several different types of alternative data that people were experimenting with at that time. You know, examples are payments on other sort of non-loan obligations like utilities or rent or even, you know, people paying for their magazine subscriptions, which used to be a bigger thing than it is today. There were people utilizing all different sorts of data on the international stage as well, looking at telecom data as an example, information that could be scraped from people’s behavior online. So we evaluated all these sorts of things. And what we found was that actually the most powerful source of new information, the best solution to this problem was kind of right underneath everyone’s noses. It was in the bank statements that were already widely incorporated into financial transactions, but that hadn’t yet been treated as a primary input into underwriting processes and hadn’t been analyzed using modern technologies like machine learning and data science.

PR: You know, I last had you on the show was back in 2019 when you were with Petal and now you, know, Prism Data, I guess, was incubated inside Petal. Is that fair to say? And then tell us a little bit about the origin story of Prism Data.

JR: Right. We latched onto this idea that as we digitize people’s bank statements, there’s a new primary source of information, just like a credit report that could be used for underwriting purposes. And that became one of the core differentiators of Petal, which is a consumer finance business that we started around that time. By 2019, we had been in market for several years with a credit product that was being underwritten largely with cashflow.

And over that period of time, we really had the opportunity to experiment and learn and evaluate the efficacy of these new approaches that we were piloting. And we found that cash flow worked really well. Not immediately. It took time and iteration and learning. By 2019, 2020, we had originated about a billion dollars in credit at Petal, largely to consumers that had thin or no credit history. And the performance of those consumers was really good. In fact, consumers with no credit score that we underwrote exclusively with cashflow had gone on to earn an average credit score of around 680, which was a prime credit score. So these were people that were prime borrowers, but were not yet acknowledged as such by the traditional credit system. That was very powerful and quite novel. Getting to that point had required us to really rethink the way that credit decisions work at the infrastructure level.

A lot of original sort of proprietary things that we built around the bank account data itself, how to clean it, categorize it, and interpret it. Then we had to think through all of the different ratios and attributes and measurements from bank data that can tell you something about someone’s credit worthiness. And then finally, you know, it was a huge amount of performance data that had to be generated. You have to, in lending, you have to actually go out and make loans and then see how they perform to know anything about the efficacy of your approach.

It was both powerful, but it was also difficult to build. And we had a growing hypothesis that we could take the same system that we had built for Petal and apply it to other loans and other contexts to help other banks and lenders to do the same thing. And so by 2020, we started this initiative internally at first to take that technology and make it available in a B2B sense to other banks and lenders. That became Prism Data. And as it gained traction, we spun it out from Petal into a fully independent company.

PR: Okay, so I’m sure you’ve had dozens, maybe over 100 conversations in the last three days just in the hallways and in the cocktail parties and what have you. When you’re first meeting someone, how do you describe Prism Data exactly? And maybe in the answer, could you also tell us sort of what kind of financial institutions you work with?

JR: Sure, one simple way to think about this is it’s sort of next generation credit scoring. And then within that, essentially what we do at Prism Data is we take bank account transactional data. So the contents of a consumer’s bank statements, we take that information and we translate it into risk scores and attributes. So we can analyze bank statements in real time and then convey what those statements tell us about the risk profile of a consumer or a small business.

Those insights that we’re able to derive from the bank account data have become highly accurate and are quite differentiated from the information that you can glean from a credit report. They tell you something different in many, many cases than what you learn from a credit report. And so, you know, when you look at our business today, the use cases that we serve are extremely diverse. We define the market mostly as anywhere that a credit score, a traditional credit score is used today, we could potentially apply cashflow underwriting to improve those decisions. And so, you think about where credit scores are incorporated, this applies to everything from very small dollar and small increments of credit risk. You can think about things like buy now, pay later, or cash advances, or even the amount of credit risk associated with a payment transaction and understanding whether or not that transaction will actually go through if the counterparty is credit worthy, all the way up to larger obligations that were involved in underwriting. So think about mainstream credit products like credit cards and personal loans and auto loans and student loan refi. The only category where I cash flow underwriting is earlier is in the mortgage space. I think that there’s great potential there, but the mortgage market is sort of slower to change in this regard because of its unique structure.

PR: So maybe you can describe kind of what goes into building your models. For cash flow underwriting, you’re creating a completely new type of model than what’s out there with traditional credit bureau state data, right?

JR: Yeah. mean, there was a generation in this industry wherein traditional credit bureaus and scores were created. Right? When you think about the trade associations that began collecting consumer loan payment information at the beginning, the formalization of that, the FCRA in 1970, then FICO and the scores they created and then popularized through the 80s and 90s. In that period of time, there was a huge amount of experimentation of creation of new structures. And then you had a period of time where we were able to rely on all of that infrastructure quite a bit. Very novel and required lots of new problems to be solved that nobody had thought about in the past. During that formative period of kind of the first generation of credit scoring, you had a common structure applied to all of the data that we consider to be credit data.

You know, those that are close to credit reporting understand the common data format. know the name of the common data format, Metro 2, right? All of the information that goes onto a credit report is reported in a consistent fashion. And then the whole industry has learned how to interpret that information and apply that to their business. None of that infrastructure exists in the world of cashflow. Similarly, when you’re making decisions with credit reports, there are a standard set of reasons required by applicable regulation that you communicate to the consumer if you decline a loan application, instance, adverse action reasons. And when making decisions based on credit report information, everybody uses the same set of reasons and the same language. All of the explainability associated with the first version of credit scoring has been made consistent. No such standard approach exists in the world of…existed in the world of cashflow underwriting. So when we were creating this in the first instance, we had to go all the way to the infrastructure layer and say, first, we need to understand all of this data that’s coming in via bank statements. We need to structure it in a consistent way.

We have to be able to explain the insights that are created from it in a way that will pass regulatory muster and that will be well understood by consumers. And all of that is, you know, perhaps even a more difficult challenge in the world of bank account transactional data for the very reason why this data is so powerful it is extremely detailed and extremely diverse think about the contents of somebody’s bank statements it includes the line by line record of what’s going on in the financial life of a consumer or what’s going on transaction by transaction with the business every cup of coffee that’s purchased every transfer from one account to the next, right? Every bill that’s paid is creating a new entry in this record. And it can be quite messy, right? Anyone who’s looked at their bank statements has seen the variability associated with this data. So a huge part of what we have to do is take in all of that semi-structured messy information and clean it up, categorize it in a way that is consistent and that is accurate.

And this is where we’ve been able to leverage some of the most sort of cutting edge statistical techniques and machine learning and AI to be able to do this well. We have to create a clear understanding of every single transaction. And then from there, we translate this data, which is now clean and clear into what we call insights. But these are attributes or features based on the banking connectivity. So it’s not important to know that Jason bought a cup of coffee this morning, if at end of the day, trying to make a credit decision. The fact that I bought a cup of coffee doesn’t matter, but if I spend $250 on coffee every month, and if that represents material portion of my budget, maybe that starts to give us some more insight into my ability to afford a new financial obligation, right? So we begin to create the measurements that are correlated with risk and ultimate financial behavior that we care about.

And you can imagine these things being, to give some examples, measurements as simple as what’s the average amount of income that someone is making on a monthly basis to things as complex as if this consumer or business takes out a loan and we can see that in their financial history, how did they spend the proceeds of that loan? Did they use it in a productive manner, maybe to pay down higher interest debt, or did they take that loan and go on vacation, buy a bunch of merchandise?

There’s a huge number of insights that we’re able to derive from the bank account data. Of course, all this happens in an automated fashion in our actual product. And then finally, we create scores. And scores for us, we have a product that we call the cash score, which is our version of something like a FICO score or a vantage score, but based entirely on bank account transactional data. And it communicates the likelihood that a borrower will default on a credit obligation based on that underlying information that we’re processing.

PR: So, if someone does that, like they’ve taken out a debt consolidation loan, but they’ve what they’ve really done is gone on vacation in Hawaii, which you can tell from the data. Does that type of thing enter into the score that you’re to provide a lender because they might have taken out a different loan that they use it in a different way than what they said they were going to use it for.

JR: Yeah, I mean, you’re just starting to scratch the surface of the amount of nuance and detail that you can get into with this sort of approach. But what matters the most to the scores, and in fact, the reason why we decided to call this form of underwriting, cashflow underwriting in the first place, is the overall financial picture that comes into view through the analysis. Right? So, you know, that measurement of how you use loan proceeds, that’s just one data point among thousands that factor into the ultimate score. What matters most at the end of the day is an understanding of the money that a borrower has coming in to repay their obligations, the large payments that they have going out that they’re obligated to make on a monthly basis, any savings or collateral that they can draw on and how all those pieces fit together. And we’re able to distill all of that into a simple score, which is highly, highly accurate in predicting a likelihood.

PR: So just staying on that for one more second, if you’re seeing a regular transfer to a savings account every month, every week, whatever it is, is that a signal that there are better potential risks?

JR: Great signal. It’s a great signal of credit worthiness. Absolutely. One of the things that I think is so powerful about cashflow underwriting is that a lot of it boils down to common sense. I’ve been talking about the nitty gritty of the data and the AI and the machine learning that’s involved, but at the core of what we’re doing, we’re talking about the basics of personal financial management. And one of the features, one of the advantages of cashflow underwriting that we’ve seen, we’ve been in market doing cash flow underwriting either ourselves or powering it for our partners since 2017. One of the things we’ve seen, there’s a very low amount of consumer complaints, questions and disputes related to credit decisions that are made with cash flow. The decisions themselves are much more intuitive than decisions that are based on credit bureau information, which can be pretty opaque and difficult for consumers to understand.

PR: Yeah, and can have mistakes. It’s a lot less likely to have a mistake in your bank account data, right?

JR: It almost never happens.

PR: You’ve recently announced some big partnerships I saw Elevate in the personal loan space, Step in digital banking, Synchrony Financial, one of the largest credit card issuers early this year. Which lender segments are actually adopting cash flow underwriting at scale right now?

JR: So cashflow underwriting started as a niche. think the earliest adopters were fintech companies like Petal that were focused on underserved categories, thin file and no file borrowers and the like. Now cashflow underwriting is becoming a core capability being used across the whole credit ecosystem. Like I said earlier, just about everywhere that a credit score has been used historically, there are now people that are adopting cashflow underwriting. The strongest momentum, you know, if I were to create a chart of the categories of lending that have the most significant adoption and then rank all of the others. I’d say the categories where we’ve seen the most rapid adoption are credit cards, personal loans, auto loans, small dollar lending, point of sale finance, including buy now pay later. There’s also beginning to be very strong uptake among depositories that are seeking better insights about their existing customers, deepen those relationships, to anticipate needs. Like I mentioned earlier, the category that probably lags the furthest behind is mortgage, but there is significant activity in that space. Both Fannie and Freddie have begun studying cashflow data and are at least learning about it as something that could potentially one day be a component of the qualified mortgage process.

PR: So it feels like 2025 has kind of been a seminal year for cash flow underwriting. The start of the year, there was some momentum, but it seems like there’s been a lot, as I mentioned earlier. But I’m curious to know, like the people that you’ve either talked to here at Money2020 or others that you’ve spoken to, what’s the reason and what do you think is going to happen to those people that just keep their head in the sand?

JR: So this is a big cultural shift and folks have become very comfortable, very adept in using credit bureau information as the primary driver of their credit decisions. So there’s quite a bit of new learning that is required and the whole industry is becoming more and more educated and engaged around the use of cashflow data. We’ve had the opportunity now to test our solutions on over 100 different loan portfolios of different banks and lenders. And so, you know, I think there was an initial question of just the efficacy of this approach. We’ve been able to now demonstrate that efficacy across a tremendous amount of data. I would say that some of the organizations that are less equipped to integrate new technology providers are falling behind a bit. So there’s less adoption so far among small credit unions and community banks, in part because these organizations often lack the engineering resources to easily bring in a new technology tool on their own. I think it’s really important that those institutions are not left behind. And one of the ironies here is that credit unions and community banks have actually been some of the foremost practitioners of cashflow underwriting historically. we think the key to unlocking these markets is to incorporate cashflow analytics into the core systems and the technology providers that are already integrated. That’s going to be an important goal for us in the coming years to ensure that we’re able to serve those markets well.

PR: So we’ve read about the predictive power of some of the traditional credit models and traditional credit scores. What do you think is the reason that we’re seeing this degradation in some of the traditional credit models?

JR: You know, traditional credit scores have never been good at adjusting to rapidly changed circumstances. One of the advantages of cashflow underwriting that we haven’t touched on yet is that it’s more or less real time. You’re able to see information from the last 24 hours of transactions. So there’s very little gap in what’s happening in the consumer’s financial life and what the lender is able to evaluate with credit reports and scores.

There is a significant period of time before new information gets furnished and incorporated into those things. We really saw this in the financial crisis first and then again during COVID and the aftermath. At the end of 2020, which was a period of time of deep disruption in the economy globally, you had millions of Americans that were out of work that couldn’t go into their jobs. For instance, FICO scores hit an all time high at that point in time.

Now lenders then were smart enough to know that people hadn’t reached a new level of credit worthiness in that moment. So they knew that credit scores weren’t telling them the whole story. But the problem was in the absence of some other reliable source of information, it left the industry flying blind and access to credit sharply declined in that period. And in the years that followed, we’ve seen credit scores become more and more divorced from reality. Chief risk officers that we talked to have seen credit scores that they’ve relied on for years and years begin to drift where a 680 FICO doesn’t mean the same thing that it did several years ago. And I think there are a few important factors that driving this. First, there’s been a huge amount of growth in forms of debt and borrowing or debt-like obligations that are not reflected on credit reports. So buy now pay later, of course, is one of the largest trend shifts in the credit market and largely not reported to the bureaus, rent to own, lease to own arrangements, small dollar credit, cash advances. The use of these sorts of products pretty widespread and not factored in to traditional credit scores and credit reports.

A lot of consumers today have far more leverage than you’d think in looking at their credit reports. So the basic debt to income ratio that you would calculate on a consumer is just incorrect. It’s just erroneous. That’s the first one. Other types of debt that are important to understanding consumers’ overall financial position, like student loan obligations and medical debt, have been suppressed from reporting to the bureaus based on various public policy interventions. And these policies have various goals in mind and there are some very good intentions at work from a consumer perspective, but by suppressing key information, you make it harder for lenders to find ground truth, which then makes it more difficult for them to make credit available. And then finally, we’re starting to see the introduction of credit optimization tools that are driven by new product structures and AI and increasingly, consumers can pay a fee to have their credit score increased.

PR: Suddenly they’re more credit worthy.

JR: Almost magically, right? Without taking any action on their part, without actually demonstrating a history of repayment. Again, it’s another example where the intentions are quite positive, but the impact is that the credit scores and reports are less useful than they used to be. So increasingly you need a more sophisticated data-driven tool to cut through all that noise and see the clear picture. We think cashflow underwriting is that tool. It’s not a nice to have anymore. It’s essential.

PR: So then how is it being used? I mean, obviously it started with primarily second looks where a borrower would go through an initial application and not meet the lender’s credit models and then go into a secondary model based on cashflow. Is that still what most companies are doing and is that going to change or how is that going to change to like a first look sort of scenario?

JR: Yeah, this is evolving and advancing quite rapidly. So to start by describing the second look use case, which is quite prevalent in the market now, a consumer applies for credit and based on their traditional credit score or credit report, they fail to qualify. They don’t meet the lender’s criteria. Instead of the lender saying no at that point, they ask the consumer for more information. And the consumer then, by the way, I’m saying consumer, but this could all relate to a business lending as well. The consumer, the applicant then has the opportunity to bring forward more financial information to demonstrate their credit worthiness. And what we do is we help the lender to understand that information. Increasingly, there’s a significant portion of applicants that can then be approved via cashflow that previously would have been, you know, flatly declined. So we see this kind of a structure as oftentimes the starting point for lenders that are just getting into cashflow underwriting. Another structure that has become more popular, especially in the last 12 to 18 months, is an offer improvement-type of process where a lender may be able to make an initial offer of credit based on their traditional approach. They start there. Hey, congratulations. You’re approved for the auto financing that you applied for. Here’s your down payment. Here’s your APR. Now, if you like, you can share more information. And if that information builds our confidence in you as a borrower, we can potentially give you a better offer, lower down payment or better APR. That’s a really great sort of win-win in that consumers that want to participate can opt into it but they’re not required to.

Now, where we’re seeing a really big shift where cashflow is beginning to come on to the same level as bureau data, and in some cases even achieve primacy to the bureau information is in non-prime thin file, no file type lending. Half of the US population has a credit score below prime. And we’re seeing that some of the largest lenders to that population are now requiring cashflow underwriting on 100 % of their applicants. So all consumers are linking their bank account. And then in some cases for products like cash advances and things like that, the credit report isn’t being consulted. And cashflow underwriting is the primary driver of the credit decision.

PR: So then one of the knocks on cash flow running has been the friction involved in connecting a bank account. Now we’ve seen improvements with Plaid and the introduction of Plaid Layer where I was doing something at connecting account the other day, just typing your phone number and boom, you’ve connected your bank account, which really, really reduces the friction. But is that still a big point of contention when you’re talking with lenders?

JR: The friction associated with moving the data from the consumer’s bank account to the lender for the underwriting evaluation has been one of the biggest historical impediments from mass adoption of cashflow underwriting. That impediment, that hurdle has been lowering really rapidly. So the first thing to understand is that cashflow underwriting is not limited to use cases where a consumer is sharing data from a third party bank account to a lender. Every depository institution that wants to lend money to their customers can apply cashflow underwriting to the existing deposit data that they already have. And that’s a very powerful use case for traditional banks, credit unions, for digital banks. For non-bank lenders, the data is coming from some third party depository. And then they’re using a platform like Plaid, for instance, to connect to the bank account and pull in that data. The conversion rate through that experience, the success rate at which a consumer is able to access and share their data has become dramatically higher, dramatically better over the last five years. Part of that has to do with better technical infrastructure.

The data aggregators are connected to many more of the banks via API. Part of that has to do with the consumer’s familiarity with the process and just the ubiquity of cashflow underwriting, so it’s kind of self-fulfilling in a way. So many consumers have connected their bank accounts in the past to access some service or to make a payment that they’re comfortable with the process. And then, you know you mentioned the Plaid Layer experience, which has really been a game changer in this category. Plaid has now seen 50 % of the U S population. And if they’ve connected a bank account for you before, they can often pre-populate that connection without you keying in your username and password once they identify you often using your bank account. And so as that friction comes down, cashflow data can be accessed in a way that’s more similar to Bureau reports. And then we get into all of the uses for that data, which are really, really manifest.

PR: So I just want to touch on briefly the state of open banking because we obviously had the plaid deal with Chase. How impactful would it be for Prism if you had to pay more for access?

JR: Cashflow underwriting is not the use case for open banking that’s really in the crosshairs of the current fight that’s playing out. A lot more of the focus is on payments, but cashflow underwriting is kind of swept up in it to a certain extent. The economic rationale for cashflow underwriting is very strong. If we’re able to enable a lender to approve a new loan that they were previously going to decline, oftentimes that loan is worth hundreds or even thousands of dollars, right? To a lender. And so I don’t think that the current negotiations and debate that’s taking place in the regulatory space, I don’t think that that poses any existential threat to cashflow underwriting. I think irrespective of the value of, the cost of the data, there still is a very strong business case for cashflow underwriting. And the business case is only getting stronger as services like Prism are able to provide more and more predictive power from the data that you’ve acquired.

PR: So the cash flow underwriting space, I mean, you’re not in this alone. It’s becoming more competitive. We’ve seen a number of different cashflow scores, cash flow underwriting platforms being offered. What sets Prism apart from all of these other people?

JR: The first one, more predictive power. We have been developing cashflow underwriting for 10 years and a lot of the other providers in the space have created models in the last year, maybe the last 24 months. And I know what those kinds of models are like because I remember when we were one or two years into developing these approaches. So in this space, you’re often able to compare the predictive power of solutions in a quantitative back test. We tend to win all of those back tests that we participate in. Today, our cash score, which is our analog to a FICO score or a Vantage score, it generally produces more predictive power on its own than a traditional credit score on a standalone basis. And you won’t hear anyone else make that claim. And I hesitate to make it as well. If I hadn’t seen this played out over dozens and dozens of back tests, because it is a very bold thing to say that this totally alternative process can generate the same amount of predictive accuracy as the credit scores that so much of our financial system is based on. That’s notable, but it’s not the typical way that cashflow is used. It’s usually used in addition to traditional bureau data where we produce about a 30 % predictive lift over and above a traditional credit score. Our competitors don’t claim a 30 % predictive lift because they’re not there yet. That’s the first thing.

The second thing is maturity. We understand all of the pieces that are necessary for cashflow to be successful outside of just the analytical component. And then finally, we built a solution that is agnostic to data source, fully flexible, that can use data, whether it comes through an aggregator, through a depository directly from a bank statement, and can produce the same output, the same score at the end of the day, irrespective of source.

PR: Well, I’ve just been given the signal here from the Money2020 crew. We have to wrap it up. We’ll have to leave it there. Jason, so great to chat with you today and best of luck.

JR: Thanks Peter. Always a pleasure.

PR: I want to go back and highlight probably the boldest thing that Jason said in this interview. That is the cashflow underwriting now matches or exceeds the predictive power of traditional credit scores. He said the Prism’s cash score produces more predictive power on its own than a traditional credit score. And when combined with Bureau data, it delivers about a 30 % predictive lift. And here’s the thing, the cashflow models keep getting better at a much faster rate than traditional models. So that is unlikely to be the ceiling here and more and more lenders are starting to recognize that fact.

Anyway, that’s it for today’s show. If you enjoy these episodes, please go ahead and subscribe, tell a friend, we’ll leave you that. Thanks so much for listening.