Blend Founder Nima Ghamsari on Customer ROI, Culture, and Vertical Software

Foreign. Hi, everyone. I'm so, so excited to have a great guest today. In conversation today, Nima Gamsari, founder and head of Blend Labs, Inc. You know, Nima, one of the most interesting thing is that I was looking at your bio and you don't say you're a CEO, founder and head of Blend Labs, which is so unusual for a public company. I want to dig it. But first of all, welcome and thanks for joining our podcast.

Yeah, thanks for having me.

So let's start at the beginning. You know, you founded the company and before that you worked at Palantir, and before that you went to Stanford. So walk us through your entrepreneurial journey. You know, did you always wanted to start a company or did you see something at Palantir? How did you come across Palantir? Right after college?

Yeah, I mean, actually, I was going to be a quant on a quant. I was a software engineer in college, and so I was going to be on a quant desk at a bank. And a week before I was supposed to start, it was before July 4th weekend, I got a call from my sister and she's like, oh, you should check out. This is 2008. She's like, you should check out this company, Palantir. And I'd never heard of them. It's a small company, 50 people or whatever. And so I go, you know, I. I take the call from somebody she used to work with, and he's like, come to the office. Just come to the office. And I'm skeptical because I'm like, I start my new job. I already had an apartment in New York. I had put a deposit down in the apartment. I had roommates I was going to live with. And I end up going and seeing the office. And there's magic gathering cards. You know what that is? Like a trading card game. Magic gathering cards all over the floor. And. But the thing that was really interesting about it, because it was such a small office and a small team, was it was everyone there was so smart. It was like the smartest people I'd ever met. And all the people from Stanford who I'd, you know, would really looked up to were there. And so I remember, I. They were like, you should really consider this. And they made me. I visited the office on a Wednesday. I did my interviews on Thursday. And then Friday before July 4th weekend, they, you know, I accepted the offer and I called my boss at. At the bank that I was going to go work at, and I was like, hey, I'm not coming in Tuesday after July 4th weekend. And he's like, oh, you're coming in like Wednesday or Thursday. And I was like, no, no, I'm just not like, I'm not starting. And I joined Palantir and Palantir is a super awesome company, always has been a super awesome company in that it was. The thing that's unique about Palantir is that it's the only. Almost all the tech companies are inside out, meaning they look at the problem from their lens and then they try to apply it to a space like a, you know, some, some vertical or some, you know, some horizontal. And they say, here's our technology. It does these things here. You take the software, you take the technology. What was so special about Palantir is that you'd work backwards from this problems that clients had and create solutions for that, which was controversial because they would, you know, people would say, oh, that's consulting. Well, right.

Yeah, that was big knock on the company.

It is, it is still a big knock on the company. But the reality is like I'm the CEO of a public company now and somebody comes to me and says, hey, I'll give you this software. I'm like, okay, I'll pay you, I don't know, a million bucks for it. But if they come to me and that software can have 100 million of impact. But like I have to do a lot of work. But if you come to me and you say I'm going to make $100 million impact to your business and all I want is $10 million, like I'll pay 10 million for the same thing. And so yeah, maybe you're putting one or two people as forward deployed engineers on that. But you know, Palantir can monetize solutions rather than. And outcomes rather than inputs. And so anyways, one of the inputs that right then the financial crisis hit and one of the inputs that I or outputs that I saw was that the financial crisis needed a lot. The banks needed a lot of help and we were working with them as Palantir at the time. And I was helping build out some solutions for large financial services firms. And the origin of Blend was I just saw how dated the technology was. I mean they were using technology that was literally the first line of code was written in the 1960s to manage $10 trillion in mortgage assets.

Wow.

That does not make any sense. That can't be right. I was like, why is nobody built this? And you know, I never actually thought of having to build a company from scratch in the sense that it was more like there was something that I was like, this really needs to exist in the world and somebody has to build it. And no one's building it because everyone's working on social media. At the time everyone worked on social media and like, you know, consumer tech. And so nobody's going to go build software for banking or for the mortgage industry. It's not the sexiest thing to do. And so we set out in 2012 and the rest is history.

And so how did you like go about starting the business? Did you find a co founder or did you like, like what triggered like obviously saw the opportunity but you know, leaving a company like talented, that's growing really fast definitely was a very difficult decision. What put you off the edge? Like, you know, I'm going to go do it.

It was a lot of factors. But yeah, I had a couple other people who were working as, you know, either engineers or forward deployed engineers at Palantir with me and we were like, this has to exist in the world. What's in Palantir wasn't going to build vertical specific software at least at the time that was like very tailor made for a vertical and it was more about a horizontal platform that could work across all industries. Like that's their DNA. And so we're like, what's the best way to make this exist in the world? It's to go start a company. And so a few of us left and started this company and you know, the Palantir team was actually super sort of supportive of that and they wanted, you know, to be a part of it either you know, emotionally or in some cases financially. And so it was actually really good and it was actually really like a fun time. We set out to do it and, and it was maybe 10,000 times harder than we thought it would be. But it's, it's been a good, It's a good 12, 13 years since then now.

So what are the like one thing that you learned at Palantir that you brought it when you started the company, one value or one strategy that you learned that said, hey, that is a very good thing I learned and that helped you build the blend.

Kind of goes back to what I said about what makes Palantir really special is they do focus on the outcome, the ROI for the customer. And one thing that most, like I said, most tech companies don't do this. They're focused on their technology and their product and maybe their sales of the product, but not the end outcome. Like even most companies that, you know, they, they, they do care about sales, they care about getting the sale done. And then they're like, okay, we got to go get another customer now, so let's go make another sale. But Palantir did teach me to work backwards and. And really drive ROI for the customer as a primary mechanism. And that happens to be. I was lucky enough to meet this guy, Peter Gastner, early on in the. The. The blender life cycle. And he's the CEO of Viva, which is a huge vertical software company. Now, they were much smaller back then in 2012. And I was asking him, I was like, what? Like, what do you think is, you know, makes Viva special? What do you think about vertical software? There aren't a lot of vertical software companies that have been built in the last 20 years. Why is that? And he was like, one, it's just super hard to build vertical software because everything's so nuanced and so detailed and so integrated. But he said, two, the best vertical software companies put most of their attention on customer success and customer roi. And he said, he's like, no matter how focused you are on your customers, it's never wrong. Like, the more focused you are, it may seem almost irrational. But that was the same lesson from Palantir, like, be super focused on your customer that Palantir had in mind. And we took the same ethos to blend. And, you know, actually, one thing, you know, we announced our earnings yesterday. And the thing I'm most proud of this year, we have, like, a few hundred customers that are financial institutions, some of them the biggest in the country. And we have effectively. Shouldn't say effectively, we had zero churn notices from those customers this year. Think about most. You know, most of them. You know, good vertical software companies might have 97% gross retention. Really good ones, good horizontal might have, like, low 90s or mid-90s, if they're really good. But, you know, you kind of expect small customers to churn or maybe want to have one or two churn. We have a zero. And that's because we genuinely care about getting really integrated, driving real ROI for our customers and making sure that the benefit that they get is really there. You know, like, it's really, really there. And even our business model is tied to that. We only charge. We don't charge when you use our system. We charge when our system helps you fund a loan or open an account. We don't charge. It's super controversial at the time. Now it's in vogue again because of all the things I talk about, Palantir and whatnot. And also people think number of seats are going down in the industry because of AI over time. And so they're like, oh, seat based models are terrible. But for the first 10 years of the company's life, everyone's like, you should be a seat based model. You should be a seat based model. It's like, no, we should not. It's not tied to the outcome for the customer. And so they're willing to pay you more for a seat for a outcome based model than a seat based model.

So this is a good way to maybe talk about, you know, Blend. You know, you talked a little bit about your vision, right? But let's explain, you know, what is your vision for Blend Labs was on the day one. And if that has changed at all, if not what Blend Lab is doing, how that it became evolved so that, you know, people who are not familiar with Blend Labs, you know, give us a flavor.

Yeah, I mean, initially we set out to modernize the technology stack of the mortgage industry. We wanted to have it, you know, to have it be something that could work really well in a mobile first world, in a digital first world and in a data first world. And I'll talk about those three. I remember on our first, on our first calls with customers, I was like, you know, most of your customers over the next five to 10 years are going to be applying for their mortgages on a mobile device. And people are like, no, no, like why would they want to do that? This is the biggest financial decision of their life. Now fast forward 10 years, 12 years. 70% of the applications that go through our our system are mobile.

Wow.

Like, or some crazy number like that that you wouldn't, you wouldn't even expect because it's like you don't even think about it though, because now mobile is so mainstream, right. That you genuinely don't think about that. Digital first. Like there was this view and people are anti mobile, but they eventually got there and then they're digital first. I just mean more like humans versus working with humans versus working with systems. As a consumer, people didn't believe that for such a big financial decision, people would be okay mostly interacting with the system and only using a human when things, you know, needed to be explained or needed to, you know, went off the rails and it's just turned out to not be right. It just turns out that humans do like working with humans when they have very, very specific things. But they don't want to be on the phone with humans for hours keying in information into a system in either side, either side of the fence. And then the third one is data driven. This one was less controversial where. And by data driven, I mean it takes a lot of data to underwrite a mortgage. You need to understand the person's income history, their credit history, their bank balances, to make sure they have enough for a down payment and reserves for the future loan payments. And it was starting with companies like Plaid, companies like these new income aggregators like ADP and whatnot. They were starting to figure out, how do we actually monetize the data that's out there in a way that drives lower friction. And so we were really the first ones to bring that data over documents approach to the mortgage industry. So that was the initial vision, was to do that for the mortgage industry. And then it sort of, I would say, expanded organically from there because, you know, as a vertical software company, you get really close to your customers if you're good. And so we got really close to our customers and they said, hey, great, you solved this project or this process for us in mortgage, and now it's so much better. And we're doing so much more volume and we're doing it at lower operational cost to us. Can we just take the same general philosophy or even the, ideally the same technology and apply it to our home equity loans? Because now we have a bunch of people who have built equity in their homes and we need to make it easy for them to get home equity loans. And then it's like, oh, well, we have these people, when we bring them on for a mortgage, we want to offer them a discount to get a checking account with us. But our checking account process, they have to go into a branch and then fill out a form. Can we bring the same technology to checking accounts? Oh, well, most. And then it's like, oh, well, on the deposit account side, now that you have that, most of the time when we open a deposit account, we open credit card for that customer. And that's the number, 60% of customers who get the deposit account with us get a credit card. Can you just build the credit card process so that we can build that into the flow as well? And then, you know, it sort of just kind of continued from there. And eventually they were like, you know, architecturally, we want to have the same DNA across all these products that we offer, all the consumer products, all the small business products that they offer, which can be digital, can be data driven, can be mobile first. And so they started working on that. Some customers started working on that with us across all their product lines. And so it was organic and so now I'd say I'd simplify our vision into saying we want to be the best way for these banks to onboard and cross sell and upsell customers across all their product lines.

Got it. So how do you make money? You said you're based on outcome, so what outcome you're driving, Are you helping them bringing lead or banks are bringing the lead? And if the banks are bringing the lead, how are you helping? What, like your contribution that banks are paying you for it and how valuable that contribution is?

Sure. So yeah, just to put it in perspective, the cost of doing a mortgage is something like $11,000 right now.

$11,000 to.

For bank or for the bank? For the bank, yeah.

Why is that much?

Well, it's extreme. One, there's a, there's a salesperson involved, so we don't really touch that part of the spend. We don't bring leads to the bank. But what we do do is for the remaining piece of the Pro. It's a super regulated industry. And so there's very specific steps you have to do in very specific parts of the process to get a loan approved according to industry standards, Fannie Mae standards, Freddie Mac standards, or Ginnie Mae standards for VA or fha. And so there's things you have to do. You have to check to make sure the person has a good income history. You have to check to make sure the person has enough in their bank account to pay for the down payment, have some reserves, it makes their first three to six payments. You have to check that their credit doesn't have any inquiries on it, that that could be adverse, could adversely impact the loan. Like if somebody went and got a credit pull at an auto dealership and they're about to get an auto loan, like you gotta flag those kinds of things. And typically the way that that's done is, you know, the industry typically will have humans manually reviewing for all these things. And it's hundreds of documents, thousands or tens of thousands of pages. Not tens of thousands, thousands of pages in some cases. And humans are reviewing it. It's expensive. And then you have to have QC processes to make sure that the humans who are reviewing it didn't miss anything. Because humans are not perfect. And then situations change across the life of the loan. So like let's say that you, you started a process and you wanted to buy a $400,000 loan or home and then, you know, you end up wanting a $500,000 loan because you end up finding a bigger house or whatever. You have to re underwrite that whole loan with that new information. And so it's a pretty technical, pretty complex process. And I think what Blend does is we help manage certain parts of that process really well. Like we'll pull in the data for your, your bank statements and we'll make sure that there's no large deposits that could be a, a loan that someone gave you that you didn't disclose. Like a personal loan, like, hey Imran, I need to buy a house. If you give me 30 grand for the down payment, I'll pay you back. Well, that's additional debt that the consumer has that we have to actually track and flag and understand why they got that money and the consumer has to attest to it. We check all the aspects of the income. Now we have a blend income product that pulls income, does the income verification in real time. So you don't need to go and read that document. We worked with the agencies Fannie and Freddie to get approval with all our data sources that hey, these are coming from the source. These are not humans looking at data. So you don't need to have the lender go check it. And so they integrated us into these day one certainty programs they have. And so yeah, so we, we help manage parts of this process and we charge, you know, call it 80 bucks alone for that prop that product. And we only charge when they close the loan. We don't charge for using the application or using the product in any way.

So if you have to think about it, right. How much, like I don't even know if it's a fair question but like how much time or money you are saving for your customers because you don't touch everything but you touch part of it.

A hundred percent. Yeah. And by the way, that's expanding over time because it's such a broad footprint. You can't build. One thing about companies that you know is you can't build everything at once, especially in a process as complex as a mortgage. And so we're taking it piece by piece. But the initial kind of core product we have, which is our blend mortgage suite, has two parts. Blend mortgage and Blend Close. We charge, you know, 80 bucks for one and 40 or 50 bucks for another. And the blend mortgage product we've done had third parties go and do research studies with our customers before and after. And you know, this, these are publicly available white papers on our website that they, that they went and did and across multiple customers. And it, you know, it saves like 600 something dollars alone, something around there. But you can find on our website Blend Close is a little different where that's to manage the closing process of a mortgage, which if you're familiar with the closing process of mortgage, you sit down with a stack of papers at a title office.

So thinks I signed. I have no idea what I signed exactly.

Well, with Blend now, which is, this is something that, that our customers are really excited about. This actually came from COVID Initially they were like, we can't close loans anymore because we can't see people in person anymore. Can we do things digitally? Turns out in certain states, with certain geographies, the time you could. And so he started building this Blend Close product. Now it's almost national that you can do these things, but the time was just a few states. But you can close loans digitally, which means you as a consumer, you get all the documents in advance. You can review them digitally, search them, find what you know, make sure everything's exactly what you want it to be. You can put it into ChatGPT and say, hey, is there something, anything I should worry about in this contract? You can download the documents, do something like that. And then when the closing comes, there's only a few documents you have to sign. We're like, hey, here's the six place you have to sign. We'll sign. We signed everything else in advance. Here's the six things you have to sign. Let's do the closing in 15 to 20 minutes. And so it makes a great closing experience. And it has a huge ROI for the lender because they can close loans earlier. They don't have to wait till the consumer is ready to meet at a title office or a notary in person. And that means that the. And we all, they all. The lender also gets what's called an electronic note, which is for the note, the instrument there. They can ship that to Fannie Mae or Freddie Mac or whoever digitally now, which saves them time and money too. Instead of waiting several days and mailing FedEx is back and forth, which is the typical way of doing things.

Got it. This is very, very interesting. So I guess, like from here, how do you grow Blend, Right? So basically automate more part of the process of mortgage, add newer product like home equity. Like, what is your vision for Blend labs for next five years or 10 years?

Yeah, I'd say there's sort of a few vectors of growth that we're seeing. When we went public, we probably had. Or when we were getting ready to go public, we probably had like 7 or 8% of the industry. Now we have a much bigger share of the industry. So getting more customers on the platform is one big piece that's super important. So we'll keep getting customers but then it's things like add ons, like blend, close that I mentioned the digital closing product is a big additional value to our customers and big additional revenue for us. And then there's, that's, that's think of that as like the revenue per unit or ARPU is going up over time. So that's another one. And then, but we're not done with even the mortgage product before I get to other products because that's the third vector. One of the things that's really beautiful. I mentioned ChatGPT as a way to analyze your, your mortgage documents before you sign them. Is it probably a good thing to do if you can do it digitally? But one of the things that is has historically been an intractable problem for the industry and we're very early on this, so I don't want to overstate it but a lot of the things that you have to do in the mortgage process are unstructured prompts. Effectively it, you know the prompt from Fannie Mae might be when you get an appraisal you have to make sure that the comps are within, call it five miles of the house that you're, or two miles or some, some distance from the house that you're, you're, you're appraising. And you have to make sure that there are three exterior photos and three interior photos of the, of the house. And so things like that are kind of intractable. I mean you could build purpose built technology for each of those components, but there's a thousand of those. And so it was like you'd be building for ages and then the rules would change and you'd have to rebuild everything. And like these there were so many approaches of rules based things that people tried that just didn't work. But with AI now it is really possible to do exactly what I just said. You give an agent, an AI agent, tools like Google Maps access and then it looks up the address and it looks up all the comps and says are these all within five miles? Yes. Good, done. Check that listing off the list there sick. Are there six photos in the document are three photos of the outside of the house and three photos of the inside of the house. Great. Check done. And so I'm particularly excited about what AI can do. But I don't want to overstate it because we're early in that, in that journey. But it's something that I think it should dramatically affect the ability for our customers, the lenders to serve consumers. And you know, we as their trusted partner because again, going back to customer success first we're, we're so well positioned to do. And so we have some early, early pilots that we're doing with big customers and I'm excited to see where that goes. But then there's the last vector which is this, this non mortgage stuff which has grown, you know, 40 plus percent year over year. It's a pretty, it's like 36% of our revenue now. People don't even know that about Blend. It's growing so fast. And it turns out that it's not just mortgage that needs modernization at these banks and lenders, it's all their consumer products. And there's not good partners or vendors for these products, unfortunately. And so again, where I see Blend long term as the most frictionless way to onboard a consumer for any of these products and intelligently offer financial solutions to consumers that drive cross sell across any of these products. So hey, we noticed you're applying for a mortgage, but did you know that your auto loan that you did with Toyota or whatever is 6% higher than our current rates here at Nema Bank? And here's a better loan that we can also do for you, simultaneously save you additional money on top of this refinance. And so there's a lot to do for the industry. I think AI will play a big role in it and Blend will hopefully be the driver of that.

So you talked about market share gain. Who do you compete with?

We probably compete with different people in different segments, honestly. Because one thing that's unique about Blend is we have one tech stack that's used at the very largest financial institutions in the country, all the way down to, you know, small lenders that do, you know, 10 loans a month. But most don't do that. Most start at the low end of the market and then they try to go upstream or they start at the high end and they build something super customizable. But then it's really bespoke technology. Effectively we were able to convince our customers to be pretty standardized on our product and we built configurations and things like that. But, but at the very top end of the segment I think we're built, we're effectively competing. It's either custom technology or in house custom technology, like those kinds of things. And then in the middle market, there's a few competitors that work with the middle market. And then at the low end of the market there's more. Like these bottoms up tools that loan officers might just download themselves. And so it's totally different. I'd say our sweet spot is really like the top 1,000 financial institutions in the country, and really even the top 200, because they have enough scale to get the benefit from really modern technology, like what we do.

Right, right, right. So maybe switching gear. Let's talk about, you know, your experience about building Blend, and what are the, like, all the entrepreneurs who might be listening to this call? What are the two or three things that you learned while building, you know, blind labs that you will advise the future? Like current entrepreneurs or the future entrepreneurs?

Sure, yeah. So. So let's start with software companies. Work backwards from the customer roi. Always, like, build that into your DNA from day one because it's not natural to people. And, like, we hire lots of people at Blend, and it's sort of in almost every case, not every case, but almost every case, that's a new motion for them to think about. Don't think about the product. Think about the outcome. And the product should help create the outcome. Not, hey, we have this product. If you build it, it's just going to create all this. This value. So they start with customer roi, customer outcome first, and then actually, that translates really well to company roi. If you build things that are making your customers more successful, they will pay you more for them over time or buy more things from you over time. It's pretty simple. So that's one. The second one I would say is I learned this lesson the hard way, which is, you know, it was tempting in, you know, the. The mid-2010s, you know, 2015-2020, to raise a bunch of money and hire up huge teams. I'm sure you have some experience with this now.

Very much so.

And on the one hand, it makes total rational sense. Like, your competitors might be raising a bunch of money and spending a bunch of money. Your internal teams are stretched. They have a lot they want to do. And so you're like, oh, this person's like, oh, I need someone to manage this. This payroll thing where I want this person to manage this part of our customer life cycle. But then it has these unintended consequences, which are, you have a lot of cooks in every kitchen.

Yep.

And you have a lot less for each individual. Like, imagine you're in a. You're in a kitchen with, like, a hundred chefs make. To make one plate. Like, each chef isn't going to have the same sense of ownership and same feeling of autonomy to make sure that plate is really good. Whereas if you Have a chef and a sous chef and they're just like attached to the hip and they're working really, really well together. The plate's going to come out better than the. The kitchen with a hundred chefs.

Yep.

And so. And I'm. Obviously it's an extreme example.

No, but it's so true.

We didn't have 50 times as many people as we needed. But I do think that I would encourage people who are building companies to be extremely thoughtful about. From personal lessons I've learned, extremely thoughtful about expanding the teams and only hire that additional individual next role one at a time when it is so painful. Like, you know, when the chef is like, hey, I need another sous chef. We have 10 times as many guests now. Make them really want it and need it so that they still feel ownership and it's not a not doesn't become this very dispersed thing. So I'd say that's another big lesson that I learned maybe the hard way.

Second one really resonate with me. I now say that I think I always tell the new founders that the biggest mistake you can do is go raise too much money and hire too many people. You know, I'm a big believer. A lot of companies die not because they don't have money, but because they have too much money. It kills the culture, it kills the speed, it kills the focus. So that is great advice that you shared.

And by the way, just because you have the money, you don't have to spend it.

Exactly.

You know, it's not burning a hole in your pocket. Save the money you happen to come across, let's say you have a quarter where you have like this massive free cash flow. You know, you're in some industry as a startup and all of a sudden you make $50 million in a quarter. Don't go and hire 10,000 people or whatever. You know, just. Just do like thought be. Keep the same principles around building and being thoughtful about spending.

I agree with you. So I want to go back to the question that I flagged at the early on. Right. I thought it was very unique. Only company that I saw that, you know, you don't. You guys don't have a C level title on your website. It's all head of finance, head of the company. Why did you have this, you know, different kind of title structure?

Actually we have a blog post this from 2018, I think.

Okay, sorry, I missed that one.

No, no, that's okay. I wanted titles to be. I think the inspiration for this was actually Bloomberg. But like, I had this thought of, like, I want titles to be really functional and extremely tied to the level and role that you're in. And not, you know, I, I saw too many companies that they would have get Salesforce. Everyone was a vp. But then the same thing on the flip side. You go to Facebook and like, there's like seven VPs and a much bigger company, and you're like, wait, why? So this title is totally, like, meaningless, essentially. And so I was like, why would I want to give out meaningless titles to people? And I will always hear these horror stories of people lobbying for titles. And people did it at, you know, my prior work experience too, and it was just a waste of energy, frankly. And I was like, we're going to have two components to our title. One that or three components. One that speaks to what function you're in. Hey, I'm in engineering. The second is your specific area of engineering or something like that, Some, like, descriptor. And then the third that ties to your level. And so we have, you know, we have leveling. Just like every company has leveling. We're gonna. You're a, you know, you're a group manager. M4. Okay, great. That means you manage, you know, a team of teams. And you have. But it was never, it was always functional. It was always purely descriptive. And I can tell you the number of times I've had a conversation with an employee about titles at Blend in the last 12 years. I can count it on one hand. Because of that, we're just like, this is our title. This is how we're going to titles. And then there was all these, by the, there's all these counter arguments to it. We're like, well, customers won't respect us if we don't have this title. If I'm not a vp, then a customer won't talk to me. That's never happened. Not once. They respect you because of your substance. They respect you because of what you're bringing to the table. From a product perspective, they respect you because of your mindset and approach and attitude. And it's never once, not once has a customer or somebody said on a call like, oh, I'm not going to be with that person. They're too junior.

That's very interesting. You know, I, I, from my previous operating experience, it was such a pain. Everybody wants fancy title. And, you know, you have to decide it. It takes so much HR energy. It's incredible.

By the way, that doesn't mean we're not paying people well. That doesn't mean we're not going to make sure people are paid market or above market. If they're really good, what it means is we're not going to give them a fancy title. I'd rather pay above market, not have a fancy title.

I'm sure you might lose a lot of sometimes employees because you don't give out fancy titles.

I have not lost out on an existing employee or not been able to recruit a single person. And in fact, it's like when people have come to me and said in a recruiting process like, hey, it's really important for me to have the COO title or it's really important for me to have XYZ title. I'll ask them why and they're like, well, you know, it's good for my resume, those kinds of things. And I'm like, it's sort of a negative indicator in some ways because they're not even thinking about the impact they're having there. I'm like, if you make this, you know, you were at Snap, Snap is now this amazing large company or if you were somebody early at Facebook, what do you think is more. And Facebook is known for having very depressed titles. Do you think someone who helped make Facebook from, you know, 1 billion to 100 billion revenue is worth less than someone who was COO of a company that never even made it to being like profitable or anything like that? Like, no, no one would think that. But people have this in their mind that this is what really matters.

This is I found so unique about it when I was looking into your company. So I think to wrap up, you know, I want to go back to AI. You know, one of the big concern in the market right now is that, hey, AI will completely disintermediate all these software companies. And you talked about how you are taking advantage of AI. So which is right? Do you think AI is beneficial for your company or do you think AI will disintermediate your company or other software businesses?

Somebody asked me on our earnings call yesterday if I think that AI will also be vertical specific or domain specific. And you know, on the one hand the general purpose models are really strong. On the other hand, speed and accuracy do really matter in these domains. And while I think there's a future in which you can have the same speed and accuracy with a generic model, general purpose model versus a domain specific approach with the right domain specific tools. Like you have to know which tools to integrate. Like I said earlier, whether it's Google Maps or something else to give the agent access to, it's I would say there's probably a three to five year window where domain specific tools are going to dominate before AI is really able to has access to all the tools in the universe and is smart enough to be as smart in any domain as anybody else in the world. Something like that. I mean I could be wrong about this, but I'll say in practice the more important piece is that if you're a large financial institution, it's going to take you a while to adopt AI. Like let's say you have a thousand people working in the mortgage department. It's not a trivial question internally of like how and where do you adopt AI and what do you do with those people? Do you use that to grow your volumes? Do you use them in another part of the bank? How do you test it to make sure it's going to work as well as the people did? How do you test it to make sure that it works as fast as the people did? And then which piece you don't take on the whole thing at once. You're taking on piece by piece. And doing those piece by piece I think will take my guess in these big regulated industries it'll take a decade for them to really adopt AI. Really adopt AI. I mean like with true, you know, to the, to the level of. I think actually today AI can do about 90 to 95% of any financial product work, maybe more. I mean maybe understating the power of AI. If it can win international math competitions, it can probably do at work in and banking. But it's just going to take time for this industry it's very regulated. You worked at big companies. Big companies tend to make methodical, thoughtful decisions. And so I think it's just going to take time.

So if it takes time and you are a long term person, you're still very early building your business, how do you position your business? So it's fine in next five, 10 years, but how do you position your business that in next 10 years you don't get disintermediated?

Yeah, no, true, that's a good point. I think and I think that the reality of the business, of what I said is going to take time is that there are going to be domain specific models that know how to do these products better than anybody else. And so I want Blend to build that domain specific model that's better than anybody else and be that tool that, that is connected to all the parts of the ecosystem, connected to the regulators, directly connected to Fannie and Freddie, directly connected to whatever other systems it needs to connect to And I want to be that domain specific, vertical, specific AI platform for these guys. And they're going to want it. They're not going to all want to go build something themselves. In fact, most of them can't.

Right. They don't have the know how.

They don't have the know how. But even if they did have the know how, culturally it's very hard to get, I'll call it spurious capital at financial institutions. Maybe at JP Morgan Chase they might invest in it because it's, they're known for investing technology. But until AI is creating a lot of ROI, they're not going to want to go put 10, 20, $50 million behind it. And they have big teams that have to run everything at these big companies. And so, so it's kind of chicken and egg. So I think they're going to want something that's industry wide. It also helps protect them where if you're using the same flavor that all the other financial institutions are using, you now have a sense a safety blanket versus having your own that then makes a mistake and now you're on the hook. And so I think that they do want this domain specific partner on AI and I think Blend is pretty well positioned to do that.

This is great. Nima. I really enjoyed the conversation and we covered so many ground. Thank you so much for taking the time and joining us.

Yeah, thanks for having me. Imran.

Thank you again for tuning in. And I really hope you enjoyed the that conversation as much as I did. Subscribe to our show on Apple or Spotify or wherever you listen to your podcast so that you don't miss an episode. See you next time. All opinions expressed on this podcast are solely those of the host and guest, are for general information purposes only and do not constitute professional investment advice.