Okay, hello everybody, and thanks for tuning in today. My name is Brendan McCarthy, I'm an analyst here at Sidoti, and I'm thrilled to welcome Vishal Garg, the founder and CEO of Better Home & Finance. Vishal, thanks for joining us today.
Thanks so much, Brendan, for having me here. I'm excited to talk about Better with you all.
Absolutely. So I've been covering the company for a little bit less than a year now. Obviously, it's received a big jump in investor interest as of late, but I really want to take the next 30 minutes or so to focus on the fundamental outlook of the company, some of the key demand drivers, and we can talk about the outlook going forward. But to start, Vishal, I think it'd be great to hear your vision for the company dating back to 2014 when you started the firm, as well as the early operating years of the company.
Totally. I think I started the company because I went through a mortgage process and lost the house that I was going to buy with my family to an all-cash bid. And I said, "Wow, how does it take 60 days for a bank to basically try to figure out three major pieces of data about my credit history, my debt-to-income ratio, and the value of the house?" And I was at that time trading mortgage-backed securities, and I could figure that out in six seconds for every loan and every mortgage-backed securities trust. I just couldn't understand the disconnect. And then I dove deeper and I found out it costs an average mortgage company $12,000, 28 people to make an 800-page PDF that we all call the closing statement. And I said, "Wow, there's got to be a better way to do this.
And if it's so hard for me, someone who is steeped in fintech and mortgages, to go get a mortgage, how hard is it for the average American?" And so I sought to make the mortgage process cheaper, faster, easier through technology, using APIs instead of people, making the process instant and enabling any homeowner, if they're buying a home, to be able to be effectively like a cash buyer, and any homeowner, if they're refinancing or going on a home equity loan, to be able to get their savings instantly rather than in months. And that's what we started Better with. And what we did from a technical perspective was recognize that what we wanted to do was effectively a three-way matching engine. So we have consumers' set of attributes, right? Your credit score, how much money you make, your payment histories, how many dependents you have.
And then you have a property which has a set of attributes. How much is it worth? When was it built? Is it a condo or single-family home? What's the lot size? And then match those two with investors who have criteria. They want to buy loans that have consumers with 740 FICO scores, living in single-family residential in these particular states with these sets of attributes. And so I sought to build a large-scale matching engine that would match consumer and property data with investor criteria for the largest tangible asset class in the world, which is residential real estate, which in the United States is worth over $30 trillion and globally over $160 trillion, and basically operated as if the internet didn't exist. And that's what we started to do. We built a machine learning-driven matching engine that automated the mortgage process.
And then more recently, we've built an AI layer on top of the machine learning to build an end-to-end platform that saves consumers and our mortgage partners' time, money, and expense and frustration in either getting a mortgage or making a mortgage for their customers.
Absolutely, and you saw a fair bit of success under this model back in 2020 and 2021, much more favorable operating environment, and you generated profitability, I believe, back in those years. Can you talk about how the business really operated back in that timeframe?
Oh, totally. I mean, we went from doing about $500 million of volume in 2016 to over $58 billion of volume in 2021, almost 100x growth. I think we were the fastest-growing fintech in America at that time. By 2021, we were doing more mortgages than B of A was doing and just a shy hair of what Chase was doing across all their branches and all their people in the United States. In 2020, when we did about $25 billion of mortgages, we generated about $800 million in revenue and about $250 million in Adjusted EBITDA. That's a really great level of profitability that we were able to generate when the rate environment was a lot more benign. Now, the challenge was that going into 2021, the bulk of our model was reliant on doing refinances for consumers.
And when rates went up by 450 basis points in a much faster way than either we or the market anticipated, a lot of that refinance volume dried up. Basically, the total addressable market for our product dried up by over 95%. And so we had to really pivot hard to build what we have today, which is a much more scalable business model powered by advances in AI and built not just from a direct-to-consumer standpoint, but also built to enable other mortgage originators, fintechs, and banks to be a participant in the mortgage business.
That's great detail. I think that's a key point for investors to understand is this was a much different business back in 2021 when you did capture a large share of volume. So you mentioned 2022, obviously the increase in interest rates presented challenges. How were you still able to really focus on growth? And you mentioned product build-out during that timeframe. What was the focus during some of those more challenging years?
Oh, totally. I think we've always been a product and tech-centric company. I think it's very different, right? Most mortgage companies in their DNA are mortgage companies. They're built by mortgage loan originators who strive to build their own business, who typically were mortgage brokers or loan officers in other companies. And so they take a very mortgage-centric mindset and a sales-centric mindset and a people-centric mindset to the problem. We've always said, "What is the problem that we're trying to solve and how can we use technology to solve that problem?" And we take it from a much more product-centric mindset, which is, "How can this product take something that is a pain point for the customer and solve it?" So in 2022 and 2023, as rates went up, we focused on a couple of things to pivot the product experience.
The first was we knew that we were going to have to change the mix from 95% refi, 5% purchase to something that was inverted, and basically, the majority of our business was purchase with the rest of it being refinanced. In order to compete on the purchase market, well, Better's always had the lowest rates through a combination of being able to pass by its savings from its technology platform and savings from its investor marketplace back to the consumer. Rates mattered less to a purchase mortgage customer. They just want the house, and so what we learned is people who are shopping for a home, they care about two things. They care about speed and certainty, getting to yes and knowing that that yes is actually good to go. They care about that a lot more than they care about the interest rate.
So we leaned very heavily, and in 2023 launched the One Day Mortgage. Now, the underpinnings of being able to take something that takes 45 days to go from a lock loan to a commitment letter at a typical mortgage company and getting that down to one day was where we had to apply technology and a parallel path processing environment. And so we really leaned hard to figure out how to do that so that we, on average today, are able to deliver a commitment letter once you upload your purchase contract and lock your loan within eight hours. 42% of our purchase commitments are actually delivered in less than 15 seconds. And all that is done through what we've built as the AI. So we've always built a machine learning-driven platform.
In true AI speak, what they call that is a supervised learning network, where basically the platform learns a set of repetitive tasks over and over again and then is able to optimize the pathways to fulfilling that task. What we were able to do when OpenAI came out with its first version of ChatGPT and introduced the GenAI models is take a lot of the things that were still requiring human judgment, where the rules were not yet codified or where the consumer was interacting with a person on our side, and start having the AI do those tasks as well. So we leaned really heavily into that very early on. It was made a lot easier for us than it was for any other mortgage company because we are the only mortgage company with its entire full-stack operating system from click to close all on one platform.
That means 12 million recorded phone calls, 6 million approved customers, 600,000 funded loan documents, 4.8 billion pieces of pages of information. All of that was completely in one place for the AI agent to learn from and to figure out and to create the same level of judgment such that today we are at feature parity between our AI loan officer, Betsy, and a typical refinance loan officer. What's more, Betsy is able to surface problems that might come up with a file and come up with solutions way faster than any typical loan officer would be able to do the math on.
And it's able to do that while knowing all of the criteria and understanding all of the requirements across the entire network of 45 different mortgage investors that we have on our platform, almost 45,000 pages of underwriting documentation and rules, and it's able to do that nearly instantly. And so that's the other big thing about what we did with the platform. We leaned really, really heavily into GenAI, utilizing our unique machine learning-driven system, Tinman, and more importantly, that data advantage that we had for learning data, which was clean learning data, all observable by our AI models inside of Tinman. And then the third thing that we've done is expand our product offering. So we leaned very heavily into launching a home equity product.
We saw others who had leaned into home equity or started home equity businesses back in 2018, 2019, start to actually be able to penetrate and that American consumers were sitting on $30 trillion plus of home equity. And while we had a product cash-out refi that touched that, cash-out refis didn't make sense for consumers that had gotten a 3% interest rate and had other debts and other things like that. What's interesting about what we did with home equity is we created, again, a multi-path platform. That is simply unable to exist outside of what we've built in Tinman because most fintechs build a single-path platform and then distribute the product at the back end through securitization.
And so the typical fintech that had gotten into the home equity space had a red light, green light type typical model of student loans or personal installment loans, not a multifactorial, multi-pathway model like what we've built inside Tinman to be able to address an investor marketplace. Now, the utility function of that to a consumer is that you get a much higher approval rate for the product and a likelihood of a lower interest rate because you're matching the consumer and the property-specific attributes across a broad cross-section of investors, investors including real estate investment trusts, mortgage investors, the GSEs, and then also banks. And that's way better than just one investor with a conduit into a securitization. So we've been able to grow that home equity business so dramatically. Last year, it grew over 250%.
We're at about a $1 billion-dollar run rate in originations on home equity, up from $100 million just two years ago. And we're growing six times faster than the other public competitor figure. And our business model doesn't require us taking any of the credit risk or prepayment risk because we're able to sell the home equity lines of credit onto the investor marketplace that we've built, mimicking what we do in the mortgage space.
Absolutely. And that's a key differentiator for the company. When you look at the balance sheet and the funding mechanism, let's talk about that in a little more detail. So this is really a balance sheet-light lending model. How can investors really think about that concept?
I think you have to think about us more like a Stripe for mortgage and home equity or a Visa or Mastercard where we're a network. On the front end are consumers or partners who have consumers on their platform who are using the platform to find ways to sell mortgages and home equity products to their customer base, and so that's very similar to merchants like on Mastercard, Visa, or Stripe. On the flip side, you have investors who are seeking to buy cash flow-producing assets secured by real estate, and so we are just the matching engine in between and the processing and fulfillment engine in between, and so we don't take credit risk. We don't take prepayment risk. We don't even take liquidity risk in that the bulk of our mortgage products, 95%, are GSE eligible.
So there's a ready market, $trillions in size in that. And then on the home equity side, we've got an investor marketplace of eight funds and banks and investors, and we're continuing to grow that. So the likelihood of us being stuck with any particular loan is super low compared to a lot of the other originators out there who are loading up loans into a warehouse line, relying on securitization, relying on rating agencies. And look, I lived through that through the credit crisis with my first fintech company, 2008, 2009, 2010. It wasn't pretty when credit risk goes up, prepayments go up, and liquidity dries up. And I think that's what we've done with our platform. And I don't think people understand that really in the equity market today.
I think when they do, our multiples are going to be very different from the ones that we trade at today, which are more like balance sheet fintech lender-type multiples.
Absolutely. I agree. I think it's an important point for investors to look at. And while we're on the topic of the balance sheet, let's talk about the restructuring that you completed last year. The balance sheet is in much better shape going forward. And now it seems like you're really positioned for growth with the recent ATM announcement in the 8-K a couple of weeks ago. Can you talk about warehouse capacity now? How can investors think about the balance sheet as a whole?
Totally. I think we really right-sized the cash structure of the company starting early this year when we retired about $375 million of debt for a cash payment of $110 million, and in doing that, we generated $265 million of positive equity to the balance sheet, which was, I think at that time, even bigger than our market cap, and if you look at our market cap today, it is one-third of our market cap that we generated in positive equity. I think what that also freed us up to do was really look at a variety of strategic options for the company and enable us to start partnering up with other fintechs who had a strategic model around using what they had built in other verticals and leveraging it to mortgage and home equity.
And then lastly, where that has left us is we currently have about $575 million of warehouse capital. I think we've announced two large deals that we've 8-Ked that alone will require us to increase that to about $2 billion to accommodate the funding demand that's going to come out of that on a monthly basis. So while we don't give guidance, I think those numbers should provide some pretty healthy color on where we expect those two partnerships and how much volume we expect those two partnerships to generate in the near future.
And what's been really great is having turned around the business, having now starting to show real growth across our mortgage products and starting to show deep partnerships with major originators, leaders in their category, like the one we announced this morning with Finance of America, which is the leading home equity and mortgage provider to the over 55 demographic. We're not partnering up with the small guys. We're partnering up with these real awesome big players to help them launch on the Tinman AI platform. We're getting a really positive reception from the street and the banks to be able to provide that capacity to grow to $2 billion a month in origination capacity for us.
Absolutely. I think that provides a great line of sight there into future volume growth. And I want to get to that as well, the B2B distribution side and the catalysts there. But let's do a deep dive into the tech stack at this point. I'll turn it over to you. And really, how can investors think about what Tinman is, how it facilitates the origination process from the borrower perspective, and then also more on the back end, how it kind of drives some of those operating efficiencies for Better?
Sure. So Tinman, I think, is unique in that it is built as a matching engine at a data field level. It is not producing a loan, then dropping it like we're used to in the old days into an Excel sheet, stratifying it up, and then selling it off to investors or putting it into a securitization. It is literally matching attributes to criteria on a singular data field level.
What that permits us to do is, as a consumer is going through the process and we're capturing more and more data about them, either through what they supply or what we get from an API pull, we're able to narrow down at specific moments in time to what is the bare minimum required to generate a pre-approval, what is the bare minimum required to generate a lock, and then what is the bare minimum required to generate a commitment letter, and then finally a closing statement. And this is not specific to any particular loan type. Tinman fundamentally can do this for any underlying financial asset that requires effectively an underwriting workflow. And we've built it for the hardest financial asset to underwrite, the consumer mortgage, and the one that has the greatest amount of variability and also the greatest amount of liquidity.
But it allows us to effectively create and launch new products much more easily than the rigid systems that exist today in mortgage land, where over 85% of the mortgage market still operates on a system that was really built in the 1990s where you can't even have more than one person at a time working in a file. Just to contextualize that, many of you guys are a little older like me, right? And you remember when you're working in a doc and it was on SharePoint and you had to tell your teammate, "Hey, can you get out of the doc and so I can put my edits in?" That's actually how 90% of the mortgage industry software platforms work. And not only are there one of them, but there are literally eight of them required to make a mortgage.
And then you have Tinman, which is a brand new tech stack where at the core architectural level, you have data field level matching. And then wherever the data is unverified and requires verification, it triggers a task that can either be fulfilled by the machine operating and getting that information directly via an API, or by the machine generating a task for a third-party service provider like a title agent or an appraiser to automatically go and fetch and then take that data back and parse it and close out the task. Or for a person, whether they're here or somewhere else employed by us or employed by someone else to go and fulfill that task. And each of these tasks and activities at the underlying level are coded with the lowest friction pathway to go and achieve fulfillment.
And they operate on a waterfall model where if the lowest friction pathway isn't available, it goes to the next pathway, it goes to the next pathway. This is what enables us to be so efficient in the manufacturing process of a mortgage, where if you look at just our last earnings deck, you can see the cost of Better to make a mortgage between sales labor and operations labor is 70% lower than that of the typical industry average if you go and look at the most recent Mortgage Bankers Association data. And I think that that's super interesting because here is where you have the building blocks of real AI in action delivering real tangible measurable results in a multi-trillion-dollar industry. So that's the underlying architecture and rules-driven logic of Tinman. On top of that, we overlay the GenAI.
So we built Betsy from a consumer sales perspective to be the world's first real AI loan officer. And in doing that, what we had to do was most AI implemented on top of a CRM system usually fails because of the hallucination rate. Particularly in financial services, you've got to get the numbers right. What makes Betsy so powerful and its ability to actually be able to go end to end, and more importantly, the calculations aren't done by the LLMs, the calculations are actually done by the calculation and rules engine inside Tinman. And so Betsy can do all the humanistic aspects that AI is so good at, taking in third-party data, helping consumers or partners understand things, providing context, but then all the calculations are actually done in Tinman.
And so unlike other LLM-driven models, Betsy is right 100% of the time on the calculations, which is what gives us the confidence to put her out there into the world doing hundreds of thousands of consumer interactions a month. The next part of what we've built on Betsy is AI processing and AI underwriting. Again, when Betsy is doing AI processing or underwriting, it's asking for the next marginal document or piece of data to fulfill. And consumers might have it in many different ways. So Betsy's then able to go and figure out, "Hey, this customer can't prove their income just using a typical W-2 because they're driving for Uber or they're a freelancer." And so automatically, we'll go and parse through that customer's bank statements, which you'll get directly via API or get via permission access, and then go and calculate income that way.
So in doing that, it's not sitting there waiting like typical mortgage industry processors. On the underwriting side, Betsy's remarkably good at detecting fraud or things that are inconsistent. And since it's storing all of the data elements and is able to reference against different data elements to figure out fraud, it's enabled to effectively underwrite in a way that's better than most typical underwriters. It's also able to surface ways for consumers to get approved that a traditional human underwriter may not remember across 45 different plus underwriting guidelines. Today, Betsy is so advanced that actually no underwriter, no human underwriter at Better is allowed to deny a loan without actually having Betsy review it as a second look. That's really powerful.
If you think about what that means, Betsy is actually a superior agent at finding solutions for consumers to be able to get approved and funded than a typical human underwriter is. Now, you take all of those things together, and what you end up getting, what matters to customers and what matters to our business partners, is a much higher approval rate on your chances of getting your customer or yourself approved, the lowest possible interest rate across a network of investors where you're optimizing for fundamentally the cost of capital, the lowest error rate in manufacturing a mortgage in the industry, 30x lower than the industry average, and this is important in an industry like mortgages where back during the credit crisis, 47% of the mortgages made were underlying faulty data and had reps and warranties issues.
And then lastly, what's really important to us is the success rate of the customer. So our delinquency rate on our mortgages, even though that we lean super hard into the underwriting criteria because we're running a matching engine, we go down as low as 580 FICO, we go as high as 50 DTI, we go as high as 97 LTV, all within the parameters of the GSE programs and the FHA and VA, even at that, our delinquency rates are one-third of that of the industry. So you have better rates and a higher chance of approval for consumers and business partners. And at the back end, a much lower default rate and a much lower error rate on manufacturing of the mortgage. This is why we think people are signing up, both consumers and partners, business partners.
And the ones signing up, we've got over $100 billion plus of volume that's already been signed up onto the platform and we're just still scratching the surface in what is a $2 trillion plus annual volume business. And I think that that's what's super exciting. It's not just what is the tech or how we made the tech or what it does. It's actually what results it tangibly provides to consumers and our business partners.
Absolutely. That is very interesting. And I think that you've talked about there's an example out there of transitioning loan officers onto the Tinman platform from some of those legacy systems. And that is the Neo powered by Better, which your company, I guess it's not technically an acquisition, but you acquired the loan officers last year. Do you want to talk about that deal and really how the impacts have been felt with Neo?
Oh, totally. I mean, we're becoming a magnet for the industry's top loan officers to come and build their business on. So rather than spend a lot of the time in the first six, seven years of our life competing against a local loan officer, much like what Amazon did with its D2C and opening it up to third-party seller marketplace, we now have local loan officers building their businesses on Tinman. They're moving from these super inefficient old-school tech stacks where it costs them $12,000 to make a loan and moving to Tinman where it costs them a fraction of that, like 70% less than that, and therefore drops more money back to their bottom line that they can invest in their customers or lowering rates for their customers or doing more deals in their local market.
And then on top of that, they're getting a better economic construct than they typically would from one of these old-school mortgage companies. And we've seen that Neo now scale up to a $200 million origination run rate. They've added about $500 million plus of new mortgage originators volume onto their platform. And we're setting up to try to grow over 100% again next year to build a really great platform for the best loan officers in the industry to come and build their business on. Again, similar to third-party seller marketplace for Amazon, but maybe more in financial services land, similar to the models that LPL and Envestnet have used to disrupt the traditional RIA space.
Absolutely. That makes sense. And so for Neo, you have cost savings, more efficient production from the loan officer perspective. Now, what are some of the benefits to Better as far as your financial results? How does that ultimately flow through to your financials?
I think one of Better's core costs as a D2C originator historically has been customer acquisition costs. So with Neo, we have zero customer acquisition costs. Our typical customer acquisition cost of $3,000 a loan goes to zero because these loan officers have local relationships already built over 10, 15, 20 years of being in business in a local market. The second thing that it does is it improves our conversion rate because we at Better for the longest time, we were almost at 6% market share if you think about how many people we pre-approve as a percentage of annual American home shoppers. But the people that we actually convert into loans is 1/20 of that amount. Where did the other people go? Well, half of them didn't buy a home, but the other 90% of those customers, they transacted with a local loan officer.
So now if we're able to identify the customers that likely need the help of someone local, someone physical, first-time home buyers, credit insecure customers, down payment insecure customers, and we're able to channel them to a local Neo loan officer, and rather than losing the business to a local loan officer, we're actually able to amplify the business that our local partners get and in doing so dramatically increase the conversion rate that we're getting for these customers, which then also improves our unit economics across the board.
Absolutely. That makes sense. And you recently announced, I think it was last week or the week prior, a new B2B platform partnership where I believe it was a non-bank mortgage originator is going to ultimately transition onto the Tinman platform. I know financial details of that are slim at the moment, but you plan to release more in the coming weeks. Can you talk about that deal or that partnership a little bit more?
Yeah. It's one of the top five mortgage companies in America. We're really psyched. We're going to help their loan officers migrate from an incumbent platform where they get 10%-15% approval rates onto Tinman, where they're going to get 50%-60% approval rates on HELOCs and HELOANs through our platform and at better unit economics. So I think that's the other thing is really it was a win-win for everyone. They have an incumbent base of customers that's millions of customers in size. And now they're marketing HELOCs and HELOANs. But again, if you're someone who has a relationship, who you know personally in a local market, and you offer them a HELOC or a HELOAN, and you get them to the table, and you only get approval rates of 10%-15%, you're not going to do that.
And so we think there's this huge opportunity to partner with local loan officers or mortgage companies that have lots of local loan officers in whatever form they are in and enable them to access the HELOC marketplace that we've built plus the product, which dramatically improves approval rates and lowers interest rates for their customers. So we have very high hopes for this partnership, and we think it's just the beginning of us penetrating that channel of partner. The other big partnership that we announced is a partnership with a large financial services company with over 50 million customers.
So if you think about that in the context of size, the other, I'll use examples, but if you go into and do a little bit of research, you'll see some of the other big banks that are out there that have 50 million customers and what their mortgage penetration rates are. And if you take that mortgage penetration rate and you just simply multiply that by 50 million, you'll get an understanding of the size and scope of that partnership and the amount of loans that we're going to be able to build up to as we deploy that partnership to all of that partner's customers. So I think those are really exciting and deep validating partnerships.
And you'll hear more about them as we officially launch them, get some customer success stories, and then are able to share details about both origination volumes, but also the names and the size and scope of the partnerships.
Absolutely. It certainly seems your potential volume catalysts there are notable with the two partnerships. And you mentioned the second partnership involves your ability to offer Tinman under more of a mortgage as a software type solution for companies. Can you talk about that initiative a little bit?
Sure, so this partner has not traditionally been in the business of brokering or making mortgages, and I think that's the case with a number of banks that left the mortgage business post the financial crisis. It's the case with a whole bunch of fintechs that have huge customer bases, but for other products who now, as their customer bases are aging, don't want to lose that relationship to your traditional big behemoth bank when that customer wants to go and get a mortgage, which is typically the type of credit that people really gravitate to and where there's a unique moment in time where a financial institution can change the existing relationship that they have with a customer or a customer's open to changing the relationships that they've had in the past with their incumbent providers.
And so we're launching a mortgage product with this partner, and we have very high hopes for what it will bring for their customers in terms of both savings and relationship stickiness compared to had they gone with someone else.
Absolutely. And what does the revenue opportunity look like for Better under that B2B partnership model? How does that ultimately flow through to your financials?
We make economics both as the fulfillment and funding agent and also as the software platform provider.
Got it. That makes sense. And when you look at these partnerships, how can you describe the pipeline looking ahead? Are you having conversations every day with other firms, or how can investors think about the pipeline there?
Yeah, so totally. Our strategy is a bit of a land and expand strategy, so there's numerous verticals where we think the Tinman AI platform does a superior job to the incumbent software providers that are all sort of like 1990s floppy disk-based platforms that have been migrated to the cloud, but underlying architecture has basically remained the same, and so when we think about that, we think, "Hey, for instance, we want to be with a fintech that has a large customer base for other loan products. Hey, we want to be with a buy now, pay later provider that has a large customer base. We want to be with a traditional mortgage company that has a large customer base. We want to be with a large servicer with a large customer base," and so on and so forth.
And in each of these, we want to capture one of the industry leaders, like the top three in the industry or in that vertical, and then from there, empower them, grow with them, and then land and expand from there. So that's what you're going to see us manifest over the coming 12 months with our strategy. And in each of these verticals, there's $hundreds of billions of annual mortgage origination, and you're going to see us tap the leader, and then from there, expand to others in the space.
That's very exciting. And we'll transition to the growth outlook here. I know we've touched on a lot of different areas, but when you look ahead, you've pointed to the goal of being or generating positive Adjusted EBITDA by the third quarter of 2026. What ultimately gets you there when you look at your results?
I think one, direct-to-consumer, particularly in the home equity business, continues to scale and scale dramatically. So I think you should see that. And the unit economics are already positive in our direct-to-consumer business, and you can see them get better as we're able to unearth more and more value for our customers and generate more positive unit economics. So that's one. Two, for our partners, Neo will continue to scale, and we're expecting that business to continue to grow rapidly in the way that it has. And that will continue to generate positive profit contribution to Better. And then three, all these partnerships that we've done will hopefully scale into the size that we think where we have deep penetration into those partners' customer bases.
And as you know, these partnerships come at margins that are more significant than D2C or the Neo business because we are talking more about software margins and AI platform margins rather than the margins typically associated with mortgage or direct-to-consumer fintech. And so with all of that, if you think about how much money we lost last quarter, and that number has continuously come down quarter on quarter on quarter, you can see a pathway where all of that starts to really glide our path down to zero, which is where we hope to be by this time next year.
Absolutely. And how does the interest rate environment really play into that outlook? I mean, I guess what variables could really cause you to really speed up that timeline or maybe even delay that timeline a little bit?
Yeah. So I'm talking about something with the current interest rate environment in place. If rates come down by 100 basis points between now or sometime next year, which is people are projecting anywhere from 50 basis points to 150 basis points of rates, you're going to go from 5 million customers being in the money for a refi to 20 million customers being in the money for a refi. Last time, we went from nearly 0% market share to 2.5% market share in refi within the space of three years. And the system is more fully built. Betsy is ready to rock and roll.
Betsy, unlike other people who have millions of customers but are staffed with 10,000 people in a call center, and those people do an average of five loans a month, and so therefore are helping 50,000 customers a month, Betsy can do a million calls all at the same time. And so I think you'll see us aggressively gain market share if and when rates come down and be very aggressive about that. And I think then there's a totally different profit picture in mind for this business.
Great. That's very exciting. We had a couple of questions flow in from attendees here. From your perspective, what's the single biggest current barrier preventing Better from becoming the dominant operating system for mortgage origination involving Tinman and Betsy?
Legacy contracts. The incumbent providers have people on three-year, five-year, eight-year type contracts with inflation escalators per seat. Can you imagine per seat pricing in the age of AI? If you're doing per seat pricing, you're automatically saying, "Let's make this the most inefficient process possible. Let's go from 28 people making a mortgage to 48 people making a mortgage." So honestly, it's these legacy contracts. And when we encounter a lot of our partners, they're like, "Wow, I would love to do this, but I've got this legacy contract. Let's talk in 2026 or 2027." And I think that's okay. There's more than enough fish that are just whose contract cycles are expiring in 2026 for us to be able to make hay. But it's legacy contracts that people are stuck in.
That makes sense. That makes sense. Well, Vishal, we really appreciate the time today. It's been very helpful. Glad we got to go through the company in more detail. I'll pass back over to you for any closing remarks.
Thanks, Brendan. We're super excited about the future. The last four years were really tough. We never gave up. We kept on investing in the tech, kept on investing in the AI, and now we're coming out of it. If we have a benign rate environment, I think we're going to scale much faster and much better than any time that we did back in 2020, 2021. We won't repeat a lot of the same mistakes. We'll be super lean and hungry, and we're really appreciative of you taking the time to help us have a forum to tell our story. Thanks.
Likewise. Looking forward to continuing to cover the company. And thank you, Stocktwits, for broadcasting the event. We appreciate everybody tuning in today. Have a great day.