Welcome to Breaking Banks, the number one global fintech radio show and podcast. I’m Brett King. And I’m Jason Henricks.

Every week since 2013, we explore the personalities, startups, innovators, and industry players driving disruption in financial services. From incumbents to unicorns and from cutting edge technology to the people using it to help create a more innovative, inclusive, and healthy financial future. I’m J.P. Nichols, and this is Breaking Banks.

Welcome back to Breaking Banks. I’m your host, Brett King. This week, we’re very fortunate to have our friends from Emerge back on the podcast, our sister podcast, Emerge Everywhere.

They’re talking this week about AI and the existential risks that we face with AI and the rewards potentially. They sit down with Ken Lin, the CEO of Credit Karma, to get his take on how AI is going to affect our future. Check it out, Emerge Everywhere on Breaking Banks.

2024 is the 20th anniversary of the Financial Health Network and the financial health movement. As we’re celebrating, we’re both reflecting on where we’ve come from and thinking about the future. So for this year’s season of Emerge Everywhere, we’re hosting conversations about the headwinds and the tailwinds that will impact our progress in the years to come.

We’re going to focus today on AI. Depending on how it develops, AI can be both a tailwind and a headwind. It can be a help and it can be a harm.

AI is an enormous topic, but I’m particularly interested in one use case, AI and financial advice. While a lot of AI conversations start with the existential big picture and then get more granular, today in my conversation with Ken Lin, founder and CEO of Credit Karma, I want to turn that formula on its head and start with the details. Credit Karma is a financial services marketplace that has built a large and loyal following by making it easy for people to check their credit score.

Now part of the broader Intuit family, Credit Karma has made a big bet on AI. And so the company is at the forefront of the journey to leverage the technology to provide customers with financial advice. I’m going to talk with Ken about the risks and rewards of the technology and whether it’s the key to unlocking personalized financial advice for all.

And Lin, welcome to Emerge Everywhere. Awesome. Thanks for having me.

So Ken, we have both been at this a long time. You founded Credit Karma in 2007, really before fintech was a thing, to make it easier for people to manage their credit scores as a way to then get personalized offers for financial products. Tell us more about where the idea for this business came from.

Yeah. So back in 2007, I was in Silicon Valley for about all the three or four years and everyone had a startup. So I felt like, my gosh, I need to think about a startup as well.

And I was going through a lot of ideas. And at the time I was working, I had my own marketing agency. And this marketing agency actually helped Prosper specifically launch.

But I also had worked with a lot of other financial services companies like Wells Fargo and Liberty Mutual. And the thing that kept going through my head was, you know, one of my first jobs out of school was I actually worked for a credit card company. I was running sort of direct mail campaigns and, you know, pre-approved offers.

And I thought to myself back in 2007, 2008, that this is really hard marketing for a particular credit quality because, you know, you would go to your traditional platforms like Google or Facebook. And if the consumer’s credit is too high, they’re not interested in your offer. If their credit is too low, then they don’t qualify for your offer.

So there’s this real sweet spot that really matters for virtually every advertiser, every financial services company out there. And if you actually knew the credit score of the consumer, you could create a lot more efficiency and you could also create a better experience for the consumer because there’s nothing more frustrating than applying for a product and getting declined. So that was sort of the problem statement that we had.

And through a lot of trial and tribulation, you know, really came down to, is there a way for us to give the credit score away for free and then create a better experience? Because, you know, I think the thing that made us successful that most people didn’t recognize was that when people are actually looking for their credit score, they’re really asking for credit, right? I mean, very few people just want to know their credit score. They really want to know their credit score because either in the near term or in the medium term, they are thinking about buying a home, taking out a credit card, taking out, you know, a model loan or something. Right.

And that’s really was the context. I think all of those things came together. Yeah, it’s pretty amazing just how much the world has changed since that insight.

Now, it’s incredibly easy to get your credit score. Every every financial provider seems to be giving away credit scores for free. And yet you had enough of a first to market advantage to have built an incredibly successful business.

And in fact, so successful that Intuit acquired the business in 2020. Can you give us a sense of the size and scope of Credit Karma today and what a typical customer looks like? Sure. So, you know, we have over 130 million consumers on our platform.

It’s mostly in the United States. We also have a presence in Canada and the UK. What’s key is, you know, we actually focus on well, we got rid of our vanity metrics a long time ago.

So what that means is you can’t have multiple accounts at Credit Karma. But, you know, along with that, it really means that we have a pretty broad swath of I’ll just I’ll say in the US of sort of, you know, consumers and American consumers specifically. So we have, you know, more than one in two millennials on our platform.

For example, we have a large percentage of, you know, Gen Z, where we tend to have fewer usage as sort of, you know, baby boomers, just because as you mature, you tend to need less access to credit, which is really, you know, where we started 17-ish years ago. But I go back to what the problem statement is for most of these consumers, which is while credit has been more ubiquitous, you know, there is always a need to learn more, to expect more from technology companies that can provide both information about yourself, but really helping you find the best opportunities to help you with your finances. And I think that’s really been our sweet spot because, you know, credit has certainly been a core tenant of our business, but the focus has always been helping people improve their financial outcomes.

And, you know, I’ve had many conversations around how do we do that at scale? And, you know, I would say that’s always been our mission. So while I often talk about, you know, the business model with investors or to the street, at the core of what we do is how do we improve the financial situation of our members? And it’s really this beautiful thing that happens in credit scores, which is, you know, our interests are actually directly aligned with our consumers’ interests, right? Because from a revenue and business model perspective, as your credit improves, the monetization opportunities improve, right? Most banks are focused on sort of the prime credit segment or as you move that credit spectrum, you know, you sort of are willing to pay more for that consumer. So, you know, we are directly incentivized in many ways to help consumers improve their credit, which I think is a really beautiful thing.

So, Ken, that’s a great segue, because the reason why I wanted to have you on today’s show is really to talk about AI and the opportunity to provide even better, more personalized, more effective financial advice beyond simply the product cell. And Intuit’s CEO, Sassan Goudarzi, made a $20 billion bet on AI back in 2019, the year before he acquired Credit Karma. And just last year, I was watching excitedly as the company unveiled its proprietary Gen AI system and watched how it was being embedded in all of Intuit’s businesses, including yours.

So I’d love to hear a little bit more about how Gen AI is changing the customer experience at Credit Karma. What does that look like today? This is what’s really exciting, right? I mean, I think a lot of trends have come and gone, and most times I kind of raise an eyebrow and say, is this a real deal? And I think Gen AI is the real deal. And one of the problems that I think we’ve recognized for a number of years is that it’s really challenging from a technology perspective to give personalized experiences, recommendation, advice to consumers at scale.

We can do it with private bankers. We could do it for the top 1% of consumers because economics work out, right? You can afford a private banker because of the dollars you might have from an assets perspective. But to do that for the 99% was really challenging because there are almost sort of a, you know, infinite number of permutes and number of advices that you can give each individual consumer.

And that’s where the power of Gen AI comes. And if anyone has played with the technology, it is just transformative in how tailored, precise, succinct if you need it to be, but also voluminous in terms of content if you want it to be. And we think that’s what’s really exciting because you can actually now have these one on one experiences where the technology understands your credit situation, your income, your debt situation, the relative interest rates that you’re paying and can give you tailored advice and have an interaction with you.

If the two sentences that I gave you as a set of contexts were not enough, you can ask probing questions. And I think that’s the interaction that we’ve always dreamed about, because in many ways we want it to be a dialogue. You know, if you think about if you had a financial planner, well, it’s a series of back and forth, right? It might be a here’s a one page report about your financial situation.

You have too much debt to income. You don’t have you’re not saving enough for retirement. But you can actually have that conversation.

You can pause that conversation, right? It doesn’t have to be at one sitting. And I should also note the other aspect of what Gen AI does is it then also takes action for you. So the context and the advice is important.

But what we also believe is fundamentally important, transformative is taking the advice or doing things for you. So it can be as simple as a, hey, pay my Chase credit card bill. Right.

It knows the credit card number. It knows the the payment address or the, you know, the routing number of that specific credit card. I think those are the things that most consumers are looking for.

That’s the hope for technology. And those are just some of the examples. Let me pause you there because I want to go deeper in literally what the use cases that you’ve started with look like.

But first. I’m still learning about the technology, as are many, and I’m assuming many of my listeners, and I want to pause for a minute on Gen AI versus other earlier permutations of AI, and specifically I’m thinking about machine learning. So machine learning is a really important tool from an underwriting perspective.

And I when I think about machine learning, I think about pattern recognition. You said earlier there are so many permutations of. Advice that people might need or behaviors or choices, I would think that machine learning has a role to play here, too.

Can you help explain where that fits, if at all, and why Gen AI is such a maybe a game changer more than machine learning might be in this in this context? Of course. So, you know, so they’re both AI, right? They’re both sort of forms of artificial intelligence. Machine learning, I think, is a great tool for what I would think about as sort of automated statistics, right, for lack of a better word.

And I think the challenge with that is, you know, machine learning is fantastic, to your point about finding pattern recognition. But to an average consumer, if I give them a statistical or, you know, an algorithm with coefficients around their probability of being a revolver or owing debt, it’s way too complicated. Where I think generative AI and in particular, LLMs are large language models is now you can actually put coherent, you know, language and conversations into the paradigm.

Right. So you can take the the machine learning algorithm around your probability of approval and actually have a human conversation around saying, hey, Ken, I think you have a 95 percent chance of being approved for this particular credit card. Right.

Before, you would just get some output and we as a product team would have to build, you know, the speech or the context to every single one of those. But now we can put it together in a very simple way, one that is nuanced to a brand that you might like or just a pattern of speech that you like. It might be very concise or again, it might be with a lot of context.

And I think that’s the game changer, because while the technology has always been around, the way to communicate the benefits and to have what I think about as as normal conversations. Right. I mean, I use another example.

You know, I think many of us have purchased a home and I know when I was buying my first home, every piece of advice I generally got would be from a friend who had bought a home before. Right. So, you know, how does interest rates work? How do points work? And, you know, what is the closing conditions? And all of these conversations that, you know, my friend might be good or he might not be good.

Right. And in this particular case, Gen AI can be, you know, very good in understanding all the technical terms, giving you very sound, objective advice, because often a lot of times the professionals are somewhat biased and that it might be your realtor. It might be the broker.

Right. They have an incentive for you to close the deal. Whereas I think in this particular situation, you can actually have someone give you all of the context and give you the specific answers that you’re looking for in the time that you want or versus here’s everything you want to know about mortgages and it’s 20 pages long and you don’t have the time or desire to read all that.

I think that’s really the transformation. Got it. Well, I want to come back to incentives.

It’s a very interesting issue as it relates to how we think AI for advice may populate the world. But first, tell us a little bit about your initial applications of Gen AI in the Credit Karma, on the Credit Karma platform. If I’m a customer today, what am I seeing? What can I participate in? We really think about it on two dimensions.

One area that we’re very focused on is we think of it as the top 20 questions that most consumers have. And if you search the Internet and you go to social media sites and general finance communities, you kind of see the top 20 questions pop up over and over again. And the commonalities, it’s very hard to find the answer that you’re looking for.

And even more so, it’s very hard to find an answer that is tailored to you, which goes back to that first point, because there’s a lot of generic information. There’s very little specific information. So one of the areas that we’re first looking at Gen AI is how do we solve that problem? How do we solve the 20 most asked questions in context to your financial situation? Right.

So one of the things that we really bring to the space in context to Gen AI and what we’re doing is the idea that we actually have the data that will make the information relevant for you. So when you ask, what’s the lowest cost way for me to borrow $10,000, for example? Well, it really depends. If you’re a super prime credit consumer who doesn’t have a lot of debt, it’s a very different answer.

And potentially, if you have a home, you might be able to get a secured home equity on a credit at, you know, prime plus something versus if you’re a subprime consumer who has a ton of bills, it might be extremely expensive. Right. And that’s one area that we believe is right for innovation, right, for the data that we have and then the advice that we’re able to provide.

The other is really around services. So and doing things for you. Right.

So in addition to you can, you know, find $10,000 of credit via a personal loan with this particular lender, we think about it as how do we use Gen AI to actually streamline the process so that dollars can show up on in your bank account in the next day? So the doing is the next aspect that we think is important. And if I think we fast forward three to five years, I think that’s going to be the key. I think the transformation isn’t just in the answers, but it’s in a world where all of us have a private financial assistant concierge, whatever you want to call it, who can actually do things for you.

Right. And I always talk about and think about most consumer problems is not a lack of knowledge. It’s too much tedium.

It’s too much of the boring things and not enough of the things that really drive us. It’s those pieces that I think in many ways can be automated that take all the drudgery out of your personal finances. So you can focus on the two or three or four things that really matter to you from a life goals perspective.

And I think historically we’ve added too much complexity, too many tactical pieces into our finances. And I think that’s where Gen AI and doing for you and giving you sound advice can can really condense it down to the things that you care about, which hopefully will help people move up the financial spectrum of mobility. What are you learning so far? I know that it’s only been it’s not even a year since you went live.

What are you learning and seeing so far? I think something like less than 20 percent of consumers have actually played with generative AI or use generative AI. So it’s still a relatively small portion of the US. So that’s one thing that we’re learning is that not everyone knows how to use it.

Not everyone knows the right ways and all of the ways in which you could use generative AI from a conversation perspective. So oftentimes prompts are really important. You need to actually suggest a couple of things that I can do to help me learn about the technology itself.

So we think that is really important. Secondarily, I think it’s all driven by data. What we have found is that the general knowledge base is out there.

So, for example, if you’re using generative AI to summarize all of the articles that you have on your site, that’s OK. I think it’s super helpful. It’s concise to be able to create the metadata or a summary of what of what content you have.

But where the rubber really meets the road is on those two dimensions that I talked about earlier, which is one, it has to be tailored. That is really important. And within that tailoring, you really have to understand the consumer, meaning some people really like all of the context.

Some people really just like, you know, can you should go and refinance that at three percent? That’s your best option, right? That is important. And again, the last piece is really the doing, because what becomes the magical moment is, yeah, it’s really cool that you can give me a tailored set of advices. But when you can start doing on people’s behalf and actually take action on the advice that you just were able to provide, that’s when we’re seeing the engagement increase.

That’s where we’re seeing the adoption increase in terms of overall usage of gen AI. So it’s really on those couple three dimensions that are really important. This show is brought to you by Alloy Labs.

As much as we love talking on the show, we believe that action is more valuable than talk. Alloy Labs is the industry leader in helping fearless bankers drive exponential growth through collaboration, exclusive partnerships and powerful network effects that give them an unfair advantage. Learn more at AlloyLabs.com. Alloy Labs, banking unbound.

So when you said that a lot of people still don’t know about this technology and what they can do with it and what they can use it, was that another way of also saying they don’t trust it? That’s a good question. I’m not sure, you know, what we have found is that when we measure how helpful is the advice, it is always in the high 60s, 70s, 80s, right, depending on the question that we’re attempting. So I don’t I’m not sure it’s actually trust.

I think it’s more of a case of not understanding, you know, like how how is this information being derived and what can I do with it? You know, I think, again, a lot of college students might have used Chats GPT, for example, to write a paper. And it’s amazing. Right.

And when you apply that, well, however, however, how many other instances of conversations of articles of advice can you apply to? And does that same level of delight happen? You know, I think that that’s what people are discovering, but it requires them to actually have a interesting question that, you know, you’re able to still delight in. And I think that’s the challenge for all the developers and all of the engineers who are working on the problem and the technology. That’s really interesting.

Back on this trust question, though, you know, it’s important for me to note that despite my interest and enthusiasm for AI as it relates to financial advice for the rest of us, there are a boatload of concerns that I and, you know, millions of others have. So let’s talk about the most obvious one, the hallucination, how do you know, how can you be sure as Credit Karma that the answers that folks are getting when they’re interacting with your platform are real? Yeah, I think this is I think I agree with you, this is the most important question. So one of the things that we do is we track all the answers and we really look at and, you know, look from a quality perspective as to how accurate are the answers that we’re providing back.

And this is where the technology is not perfect, right? As we all know, hallucinations can happen. And, you know, it’s a it’s a it’s got an interesting word for basically errors. That would be that would be a euphemism.

Yes, exactly. But it’s where, you know, Gen AI is just incorrect and it just made something up. That’s absolutely true.

Untrue. And I think this is where we really need balance. And this is where the technology needs to sort of evolve.

Right. And I think this is where we’re all focused. So how do we solve the problem? I mean, we’re making great advances each and every day, but what we’re focused on is putting guardrails around the technology.

So areas that when we get to, example, for the specifics of an offer, well, you know, we don’t want some old model that was developed or LLM that was developed two years ago, right, to provide the interest rate and, you know, the terms of that particular offer, because we know that it’s outdated. So there are sort of basic blocking and tackling things that you can do to refine the answers themselves. You have to go and then monitor those answers to ensure that you have a high bar and a high level of QA associated with it.

And then this is where the innovation comes in, which is, you know, I think a lot of companies are looking at novel ways to ensure that this doesn’t happen. Now, with all that said, I think we all recognize that, you know, there are certainly pitfalls in technology, but, you know, I think with like many technologies, it’s not perfect, but the benefits that we are able to provide to the many can outweigh the costs, you know, when it is wrong to the few. But we also have to recognize that we have to do better on the few instances where it does go wrong.

A colleague of mine recently said that. We shouldn’t worry about hallucinations in the long term because they’re just an engineering problem that the engineers will ultimately fix as a as a non-engineer, that felt a little bit hocus pocus, and this is someone I have deep respect for. What do you think about that? I think I think that’s generally right.

I mean, you know, we will solve these problems, right? I mean, oftentimes the hallucinations are a combination of either sort of gaps in the knowledge base. So we don’t know the answer. So therefore, the system just kind of makes up an answer or it is, you know, a lack of data going back to this idea.

So so both of those are actually solvable, fixable. So it’s only through the permutes of, wow, we’ve actually every time we ask questions in this context, the LLM makes an error. Well, we know how to go back to fix that particular context.

So it’s just like any other engineering problem, as you you stated, right? These are not intractable errors. I think sometimes we, you know, in many ways, I think we actually think of it as actual AI. I’m not sure it’s actual artificial intelligence.

It’s not thinking on its own. Right. You look at what Gen AI is doing is effectively scanning all of the knowledge base that we have as human beings generally on the Internet.

But now we’re getting into pictures and videos and, you know, even books, right? Because all of those are being translated and turned into digital. So it’s looking for the patterns and, you know, the the answers in that context. And, you know, the downside, right.

Going back to your point about the the ills of AI is all the biases that we have in the text, all of the discrimination that we have and that, you know, in sort of, you know, the human consciousness that also exists and what it’s what’s going to require from an engineering perspective is for us to weed out filters on that and fix those specific issues. And that will take time, but it’s not intractable. Yeah, so you went exactly where I was going next, which is this issue around bias and harms.

And that can take a number of forms, but for me, it goes back to like the data that the bottle was trained on in the first place, right? Is it the data is the world as it is, as opposed to the world we would want it to be? This is particularly an issue with credit data, right? Credit data is the world as it is. And it feels like this Gordian knot or this like intractable chicken egg situation. How are you all dealing with that? And maybe you can say a little bit more about what your model, the data your model was trained on in the in the first place and how representative you are.

One hundred and thirty million. I’m assuming it was data on your one hundred and thirty million users. You know, how representative are they? How do you deal with this bias issue? Yeah, I mean, I’ll probably speak in more generalities because at one hundred and thirty million, I think the generalities are probably just as relevant as the specifics.

But, you know, this is this is a problem that’s been known in modeling for a long time, which is, you know, to your point, all the biases that exist or even, you know, the unscored area. So, for example, right, if you only provide credit to people who are, let’s just say 700 to 750, well, you don’t have any data on the performance of people who are 650 to 700, nor do you have any data on the on the people who are 750 to 800. And now if you expand that out and generalize that, you can think about that across, you know, communities around ages, around ethnicities.

And by the way, that’s that’s illegal. We can’t do those things. Right.

But more to speak to the point of there are sort of knowledge. There are gaps in our data and our awareness. So so my thinking on this is one is we have to be extremely diligent in understanding where we have strengths and where the models work.

Secondarily is we have to expand those models to be more inclusive. We have to reduce the biases. And the way that we do it, it goes back to an engineering problem.

And, you know, this hasn’t happened yet, but my hope is that the technology makes us more efficient so that we free up cycles to look into these other areas. So so an example of or to follow the analog I just walked through. Right.

If we make the underwriting of 700 to 750 so efficient that we as humans and we have bandwidth to go and understand, well, what is underwriting at 650 to 750 look like and and and so on. Right. We can expand the knowledge.

And that, I think, is the way that we can bring every along everyone along so that not only the people in the middle of the of the target benefit from the technology, but everyone can benefit from the technology. But but but I think this is a really important topic because I think compliance, diligence and care are paramount in this category. Because we know that a lot of people, underrepresented populations have been disproportionately affected by the pace of technology.

And I think it’s paramount that we don’t allow that to happen as we build this technology, that we bring everyone along and we spend the time and effort to do that because, you know, sort of in the history of finance, we have seen this. We have seen redlining. We have seen that certain people are much more apt to get mortgages and affordability for homes.

And that leads to an upward mobility and leaves a lot of other people behind. And I don’t think we can leverage the technology to say we have to be very careful that the technology doesn’t do that again. And do you have any personal experience with bias in our financial system? Well, yeah, I mean, I well, both I mean, I both as a statistician, but actually, you know, in my own personal lifestyle, I mean, we’re going to go personal on this.

I mean, you know, I grew up I came I was an immigrant from China. My parents moved to the States when I was four and a half years old. And, you know, I remember firsthand I was the translator for my parents as they were trying to get into the banking system.

And, you know, I think between the combination of that experience, but also what I know about the banking system, you know, it’s it’s easy for me to notice how hard that was, right, how hard it was to get into and understand what a banking account did, the sort of the lack of access, the lack of of availability, particularly if you’re in a underserved neighborhood. And, you know, with my knowledge today, I know that that access, you know, is empowering. It allows for upward mobility.

Right. And that simple lack of access, whether it be in the inner city on an Indian reservation, right, that prevents people from having sort of the financial gain. So I saw it firsthand when I was a kid.

And I remember how hard it was to understand the terms. I remember how hard it was to help my parents open up a bank account. And luckily, we got through all of that.

But, you know, I think it is sort of an important aspect and in many ways has shaped how I thought about Credit Karma, because, you know, we represent one hundred and thirty million consumers and a lot of those are still underrepresented when it comes to access to the financial services space. So that’s an important aspect of how we think about business, but also how we apply our technology. Yeah, I appreciate you sharing that.

So there’s been a lot of conversation of late around AI governance. The president put out an executive order on AI writ large on regulation and all kinds of other things. But governance, I think, is is a good umbrella for.

Some of the most critical questions that remain is the technology in charge, are we in charge? Are companies like you that are putting the technology to use in charge? Share your thoughts about how as a nation we should be thinking about issues of governance and making sure that that the computers aren’t in the driver’s seat. Yeah, my thinking on this was probably evolved a little bit, right? I think there was a time when I thought the greater good was more important. But now when I look back at, you know, what I would say sort of the unequal treatment of groups, I’ve sort of come more into thinking that this note of governance and the way that we use the technology is actually becoming more important, right? Because I think it’s a little bit too easy sometimes for us to say, well, look at all the good that we did at the small cost of harm that we did to the few.

But I think if you amplify that over generations or groups and over time and you use the power of compound interest, right, or the effects of what that does to groups, you realize that that’s sort of the story that’s been happening for a long time. So my own evolution and thinking in this particular area is that it fundamentally is important, right? I don’t think as an industry we should say that look at the greater good that we’ve created at the sort of harm for this group and therefore justify, you know, the ends by any means. And I don’t think we’re there yet in that we can rethink this, right? We can be in a world where we can do this responsibly and we can ensure that there’s, you know, equal treatments for every group.

And, you know, oftentimes I always think that, you know, like if you really think about the financial services sector and the system, the reality is like the top 80 percent, they’re fine. I mean, you know, they have access. It might be a little bit hard and we want to solve it for the 80 percent, but it’s that bottom 20 percent that underbanked population, that group that’s underserved, they’re the ones that can most benefit from it.

And oftentimes they’re the group that’s ignored. So in many ways, I wonder if we can actually fit the paradigm a little bit more and use the technology to help bring that group along faster, because maybe that is the group that has been underserved for the longest amount of time. What does that mean and how do we apply that technology? I think that’s the work to be done, but I’m hopeful and optimistic that, you know, a lot of companies will see that, wow, we can actually serve this group at scale in a way that we haven’t been able to do the best.

That’s an incredibly nuanced and helpful perspective and one that I actually haven’t heard on this topic yet. So I really appreciate it, especially the part at the end about, hmm, maybe we should be focusing on the people who could most benefit from this technology. We know, though, this is true in anything, right, that the way capitalism works in this country, we’re going to go to the use cases that are going to make me the most money with the least cost first and hope it trickles down as the other way around.

At the Financial Health Network, we’re hoping to, like you said, flip the script. And we’re thinking about how we can actually make sure we are focusing on the technology on the bottom 20 percent, where the technology itself helps make the business case, helps make the economics work. So I’m really excited about that.

But let’s talk now about over the next five to 10 years. You said in a year you thought we were going to solve some of these basic questions about, basic questions that people have about their finances, which is like incredibly exciting and scary all at the same time. And how do you think this is going to continue to evolve, right, Intuit and Credit Karma has a massive head start, right? You’ve been working on this for at least five years now, massive investment, and you have tremendous amount of data on which you are able to train and continue to learn.

And if I understand AI correctly, like the earlier you get in and the more data you have, the faster you can learn, right? You can kind of like compound interest, right? You can get way ahead of your competitors. Then there are the big banks. They have lots of data.

They have tools like Erica at Bank of America or Watson at, no, at Fargo, excuse me, at Wells Fargo, which is newer and et cetera, et cetera. You know, how fast can they build on those tools? Then you talked about at the end of the day, maybe we all are going to have, or we need our own personal bot, our own personal finance bot that isn’t tied to any particular company, right? That’s kind of like the meta bot. My head is already spinning from all of this conjecturing.

Help me think about how this could roll out over the next five to 10 years and what actually is going to be best for people. Yeah, I think that, you know, I think you’re right in painting the landscape the way it is, right? And I think there are both structural advantages to some of the incumbents in the space, but there are also disadvantages, right? Because if you think about things like risk-taking, their own internal biases, meaning they’re going to build services for their existing customers and not for potential customers. So I think in the next year or two, some of the basic problems like, you know, payments, balancing, general recommendations, and even bespoke recommendations will be very simple and will be available.

What bills to pay, how to build pays on time, paying the bills with the lowest interest first. Those are relatively straightforward. We’ll have a great sense of that.

In terms of the landscape, I mean, I think what big companies should be worried about is that, you know, the cost of entry is so low these days, right? I mean, technology in Silicon Valley has been a reflection of that. You know, 20 years ago, 30 years ago, it took you, you know, $50 to $100 million to build a technology company in terms of the servers and, you know, the code and the engineers. Today, you know, somebody probably could start an AI company using AI to write the code, you know, with a programmer and, you know, maybe a $50 license, right? And I think that’s where a lot of the potential disruption can come from.

But more importantly, and back to your point about where I see the future, I think it’s unlikely that we’ll have a financial services company, AI slash agent that is dominant in this space. The one thing that I’ve seen is that, you know, consumers crave choice. You know, we crave someone who is going to give us unbiased advice.

And I think that’s really where the space is going to evolve, I mean, in many ways, and this is my own bias, and this is that one of the things that we observed with Credit Karma, and a lot of people ask us, well, why is Credit Karma successful? You mentioned this in the beginning, which is, you know, like credit scores are ubiquitous now, every bank has a credit score offering. But what we have seen is that consumers who are on our platform don’t necessarily want the product that the bank recommends, because they know they’re only recommending their own products. And that, you know, it might and it might be a very good product, but at the end of the day, consumers want to be able to compare, they want the freedom to do that.

Now with that said, I think there will be a consolidation, I don’t think we’re all going to have, you know, 50 agents each. I think we might have, you know, a financial agent, we might have, you know, sort of a travel, you know, one, and we might have a health one, right, there’s probably like four or five categories of things that will be really important. And I think there’s an underlying reason for this, which is all of these things require data about yourself.

So you have to have trust in that agent, right? And only through that data and that trust, we’ll be able to give you advice. And, you know, I know, personally, I don’t want my health information with, you know, hundreds of providers, nor do I want my financial information with 100 providers. So I think that’s why there’s going to be ultimately, you know, a consolidation of maybe two or three players in each category versus every bank having, you know, a financial agent that will help consumers.

So Ken, you mentioned that, you know, there are like roughly top 20 questions that are kind of consistent or that people have about their finances. When you think about these next five to 10 years, which of, and maybe it’s shorter, which of those questions most motivates you or do you think could have the most impact for the most people? What do you think, what are the couple, one or two questions we really should be focusing on trying to help people answer? Oh, that’s a good question. You know, I oftentimes, so let me put it into a category versus the specific question, right? I mean, the one that I think is most motivating for me is roughly the 40 to 50% of consumers who are living paycheck to paycheck.

I think for that group, let’s say a meaningful percentage, let’s say half of that group can actually get ahead with the right financial advice. Right. And, you know, I think to be very clear about, I think the other half is they’re, you know, they’re living above their means.

They just don’t have the income to support either the debt load that they’ve created or the situation that they’ve been in. And many times I should note that it might be to medical or things that are outside of their control. Yeah.

But for the other 50%, I do think it’s about awareness. Right. So if you’re paying 30% on debt that you could pay 15% on, that makes a big difference, particularly as we think about that compounding.

So if I had to hone in on one group, I would say that would be the group that I think the technology, the awareness they’re doing for you can fundamentally transform. Right. I mean, I know certainly there are other groups and, you know, everyone needs help.

But I think that group in particular could benefit the most from the technology. And I would say that’s where I would love to see our time and effort being focused, because that is just an awareness problem. That is not a means problem.

That is an awareness that is, you know, ability to do something about it. They have all of that. It’s, you know, it’s hard.

It’s hard when you don’t have the dollars coming in to pay off your debt. It’s hard to get ahead in that context. That problem also needs to be solved.

But here, I think it’s just squarely in the, if you knew the information, you took the actions, you would be, you know, out of debt in five years. And then, you know, thinking about retirement. I think that’s a powerful idea and one that, you know, whether it’s Credit Karma or somebody else should be pursuing.

Yeah. Ken, Lynn, thank you so much for joining me on Emerge Everywhere. My pleasure.

That’s it for another week of the world’s number one fintech podcast and radio show, Breaking Banks. This episode was produced by our US-based production team, including producer Lisbeth Severins, audio engineer Kevin Hirsham, with social media support from Carlo Navarro and Sylvie Johnson. If you like this episode, don’t forget to tweet it out or post it on your favorite social media.

Or leave us a five-star review on iTunes, Google Podcasts, Facebook, or wherever it is that you listen to. our show. Those actions help other people find our podcast and in return, that helps us build an audience that can be supported by sponsorship so we can continue to provide you with our award-winning content every week.

Thanks again for joining us. We’ll see you on Breaking Banks next week.