Better Home & Finance Holding Company (BETR) Earnings Call Transcript & Summary

August 7, 2025

NASDAQ US Financials Financial Services earnings 57 min

Earnings Call Speaker Segments

Tarek Afifi

executive
#1

Thanks for joining Better Home & Finance Holding Company's call for our second quarter results. I'm Tarek Afifi, and I work in Corporate Finance here at Better. With me today are our Founder and Chief Executive Officer, Vishal Garg, and our Chief Financial Officer, Kevin Ryan. Besides this call, I'd ask you to take a look at the earnings release for the second quarter, which you can find on the Investor Relations site. An investor presentation is posted there as well. Some of what we say today might be forward-looking statements under federal securities laws, built on current expectations and assumptions. Those expectations and assumptions carry risks, uncertainties, and other factors described in more detail in our SEC filings, and they could make our actual results differ materially from our historical results. We take on no obligation to update forward-looking statements except where the law requires it. During today's discussion, management will talk about certain non-GAAP financial measures, which we think matter when judging the company's financial performance. These non-GAAP financial measures shouldn't be viewed as substitutes for our GAAP results and should be read alongside them; they are reconciled to GAAP financial measures in today's earnings release and investor presentation, both of which can be found in the Investor Relations section of Better's website and when filed in our quarterly report on Form 10-Q filed with the SEC. Figures described as of and for the quarter ended June 30, 2025, are a preliminary estimate as of the date of this earnings release and could be revised when we file our quarterly report on Form 10-Q with the SEC. Additional information as of and for the quarter ended June 30, 2025, will come when we file our quarterly report on Form 10-Q with the SEC. I'll hand the call over to Vishal now.

Vishal Garg

executive
#2

Thanks, Tarek, and welcome to all of you on the line for our Q2 2025 earnings call. We appreciate your presence today and the continued support as we work toward our goal of improving homeownership—making it quicker and simpler for the people we serve—by building an AI-native technology platform that reshapes the entire homeownership experience, enabling consumers to get through the complete mortgage and home equity process in as little as 1 day; we continue making strides toward that vision in which every customer can instantly handle buying, selling, refinancing, insuring, and improving their home online at a competitive price. So, the four goals we've laid out are these. First, we push hard into AI-driven growth so that both volume and revenue go up. Second, we keep getting more efficient, and that comes from the technology continuing to advance. Third, we broaden what we sell, how we sell it, and the channels we sell through. And fourth, we bring down corporate costs relative to revenue — everything aimed at reaching profitability. Better is the first scaled mortgage tech platform, and it was built to empower consumers, and more recently to empower local mortgage bankers and financial institutions through Tinman technology so they can meet their own end customers' needs. That's what makes us different from everyone else in mortgage, and even when the market is tough, it opens up a huge greenfield opportunity for Better. Our goal is still to reach profitability, and we're doing that by pushing further into Tinman technology and AI. Betsy handled roughly 600,000 consumer interactions during Q2. AI underwriting has climbed past 43% of locked loans, and we see a clear route to 75% before long. Meanwhile, our loan officer productivity, measured as funds per month, has risen to more than 3x the mortgage industry median. For a while now, we've been discussing how we get to profitability in the medium term. Thanks to everything we've achieved with our AI platform, and the momentum we're seeing as we add not only mortgage advisers who are local market professionals but also bigger enterprises looking to get into the mortgage business through what we've built, I'm happy to report that we can now see a clear path and expect to reach adjusted EBITDA breakeven in Q3 of 2026 — essentially within the coming year. Kevin is going to cover this in more detail in a moment. The platform we've created puts AI first. Although it takes time for consumers to embrace AI, adoption is growing exponentially. Among U.S. mortgage lenders, we're among the few—perhaps the only one—offering a complete technology stack consolidated in a single location, unified into one workflow, and fully API-enabled for Agentic AI. Betsy was constructed entirely in-house, drawing on our knowledge base, which has been built from more than 12 million recorded phone conversations, over 500,000 funded loan files—complete with the full consumer financial graph and property graph, aligned to investors' detailed line-item, fact-based requirements—and 6 million pre-applications accumulated across the last 8 years. Every day, it keeps learning from each customer interaction. Ever since Betsy AI went live, the share of leads that convert to locks has risen by more than 30%, climbing from 3.3% to 4.4%, and that is hugely significant for generating additional volume and revenue while extracting more profitability from every single loan. As you can see, there is still a very considerable distance left to travel when it comes to further raising that lead-to-lock conversion rate. However, because Betsy runs at close to zero marginal cost, given that everything sits on Tinman, our own proprietary internal platform, we think we can keep enhancing our unit economics as the mortgage market settles down over time, with fairly minimal added fixed CapEx. Betsy has also produced a stronger customer experience. If someone would rather speak with a human being, Betsy helps match that customer with whichever member of our sales team is best suited to them, picking based on what that person specifically needs. Our Net Promoter Score has risen from 39 up to 64 since Betsy came on board, putting us in the same territory as companies like Google and Apple, and far above where traditional mortgage lenders and financial institutions tend to land. Betsy operates on top of the pricing and eligibility engine we built with machine learning, which we call DE internally, and that engine pulls from both the consumer's financial graph and the property graph, with each individual piece of data being checked against the criteria that every mortgage investor configures straight into our platform. It can execute hundreds of thousands of pricing permutations spanning the more than 45 mortgage investors now purchasing the loans on our platform as those loans get funded one at a time. Among these loan purchases are GSEs, banks, REITs, and a growing number of private credit funds, each bringing distinct guidelines, risk-return preferences, and other criteria onto our platform. Together they account for more than $1 trillion in demand, placing bids daily on loans that fit the criteria and portfolio needs, and our platform satisfies those criteria at an error rate far below what the industry sees. Looking ahead over the coming 1 to 2 years, we anticipate substantial growth in our investor network, especially as private credit pushes further into consumer asset-based finance. Inside consumer lending, mortgage and home equity make up more than 85% of the total addressable market and are especially appealing to private credit, since this area has a strong track record during recessions when measured against unsecured consumer debt such as personal loans, credit cards, and BNPL loans. The quality of our credit and underwriting is unmatched. That said, other than that short window when loans sit with us awaiting settlement, we don't bear any credit risk at all, because the moment a loan is locked and funded, we've got forward commitments in place to sell it to investors. That way of doing things remains unlike anything else in fintech, since a huge number of originators that started out as marketplaces have by now moved toward holding credit risk themselves. Better is simply doing the same thing it has done since it launched in January of 2016. The assets we originate aren't ones where we carry credit risk. If a recessionary environment comes along, we think that will work in our favor. Looking ahead to the back half of 2025 and past that point, our strategic priorities stay centered on the things within our control. Priority number one is to keep pushing growth opportunities forward regardless of what happens in the wider economy or the mortgage market. During Q2 of 2025, compared with the same period a year earlier, funded loan volume rose 25%, reaching $1.2 billion, while revenue climbed 37% to $44.1 million, powered by funding a greater number of loans across both our D2C channel and our Tinman AI platform channel, at a higher gain on sale than in the prior-year period. During that quarter, the increase in funded loan volume compared with the same period a year earlier was driven by HELOC and home equity loans, which rose 166%; refinance loan volume, which climbed 109%; and purchase loan volume, which gained 1%. The overall growth recorded in the quarter reflects the strategic investments we have made in technology, product innovation, and distribution expansion. Those include Betsy AI; our Tinman AI platform strategy with NEO powered by Better, which is being proven out; and the efficient expansion of D2C, driven by the efficiencies we are realizing from implementing AI throughout the entire platform.

Vishal Garg

executive
#3

Through these strategic efforts, we're now set up to seize a wider range of opportunities in the market, boost how efficiently we operate, and keep generating growth that lasts, with revenue increasing faster than loan origination volume—both during the quarter we just finished and in the quarters we anticipate ahead. The second thing we're focused on as we move toward profitability is trimming our costs further and making our operations more efficient, aiming to hit adjusted EBITDA breakeven by Q3 2026, which is basically in the next 12 months. Thanks to our Tinman AI platform, we've managed to automate parts of the mortgage process that eat up lots of time and labor, and we've steadily cut our cost to originate down to roughly half of what the industry averages. In the next 12 months, we're looking to push that down even more, getting it to about 1/3 of the industry average, and then keep improving as consumers adopt AI more and interact with it more, which helps us remove costs from the mortgage manufacturing process. Even though we anticipate loan origination expenses will go up as our volume grows, we think our ongoing investments in AI—with our product and engineering road maps right on schedule—will bring costs down a lot more, leading to better operating efficiency and unit economics. Finally, the third thing we're prioritizing is expanding how we distribute our products and platform across different channels. Right now we serve the end consumer both via our direct-to-consumer model and through our AI platform model, which covers Tinman as a platform and Tinman as a software. In our D2C business, we work directly with the consumer on better.com. We started out with the mission of transforming how consumers experience the home finance process, so our D2C business has consistently led the way in stretching what technology can do in the mortgage industry. We've funded over $100 billion in loans through the D2C platform, and that's given us the foundation to train the AI and keeps creating a positive feedback loop that gets better every single day. Our D2C business gives us the chance to launch and try out new AI features at a scale and pace that no one else can really keep up with. And in our D2C unit channel, our unit economics keep getting better as we keep pushing AI through the whole process. As you can see in the earnings deck we've shared alongside our materials, our contribution margin or per loan profitability has kept rising while the operating cost of fund has kept falling, thanks to both conversion gains and putting AI into the sales and operations workflows. To be more specific, in Q2 2025 for D2C, revenue per loan was $78.86 per loan. Our cost per fund was $68.22, for a contribution profit of $1,064 and a contribution margin of 13%. Once again, we've kept fine-tuning our pricing so we stay competitive with the big names in the market. Yet, as you can see, the total expense we incur to originate a loan is roughly 50% of what the mortgage industry typically spends, and that figure already accounts for what we pay to acquire customers. We also anticipate pushing it lower and lower, since we're simultaneously boosting conversion rates and reducing CAC, while also enhancing labor expenses—which in turn should cut funding costs dramatically. Then, we take care of the customer via our Tinman AI platform, which supports neighborhood retail loan officers throughout America, and there we keep witnessing fast early-stage expansion. Under this approach, we essentially act as a platform supplier and a P&L collaborator within the mortgage origination workflow, with customer acquisition cost at almost 0, since the mortgage retail loan officers supply their own clients, their own connections, and their own deals to our Tinman AI platform. We're swiftly upending the conventional retail mortgage origination sector by bringing loan officers and branches onto our Tinman AI platform, giving them the ability to close more loans than ever in their careers, eliminating obstacles in their fulfillment workflow, and broadening their ability to assist additional customers via the lead flow that we produce from. Over time, we anticipate these loan officers can shrink an astonishing 80% of their back-office expenses by leveraging our platform. The opportunity for the Tinman AI platform business is enormous, and we've only just begun. To put it in perspective, more than $1.2 trillion in mortgage volume during 2024 came from retail loan officers and mortgage brokers using outdated technology and carrying heavy operating expenses. Merely 1% of that market would equal $12 billion in fresh loan volume, again with customer acquisition cost at almost 0 to better; we keep achieving excellent momentum on the Tinman AI platform through our initial and by now thoroughly validated rollout of NEO Powered by Better. We kicked off production with NEO at the beginning of 2025 and hold ambitious goals for what lies ahead. During Q2, we funded 429 million loans for 1,009 families using NEO Powered by Better, which is a rise of 164% and 176%, respectively, versus the previous quarter. The unit economics of our Tinman AI platform are pretty robust. To be precise, in Q2 2025, for each loan funded on the Tinman AI platform via NEO Powered by Better, we brought in a contribution profit of $6,172 on revenue per loan of $15,538, leading to a contribution margin of 40% for the Tinman AI platform. As we manage to boost adoption of the Tinman AI platform's processes among the retail loan officers, we anticipate that margin will grow even higher, letting us either pay these retail loan officers more generously or assist in generating greater profits for them and their branches down the line, which we think, taken together, forms something like a holy trifecta. The Tinman AI platform lets retail loan officers close more loans, help more customers while putting in the same hours, and with a reduced cost to fund, producing huge profitability gains for their business compared to operating on a conventional retail platform or mortgage broker platform. This combination of advantages is what lets us bring on more local loan officers who come from top-performing retail mortgage firms — loanDepot, Nationwide, and Movement Mortgage are just a few examples. In the quarter that just ended, we brought aboard loan officers from those organizations who together originated more than $180 million in volume during the prior year. And looking ahead, we anticipate drawing in even more of these high-producing mortgage professionals in the retail channel, since our platform gives them technology and a model that are clearly better. On top of that, the niche expertise these loan officers bring is extending our footprint into additional loan categories that tend to be more intricate or specialized yet carry better margins. Among those categories are nonconventional FHA, VA, and jumbo loans. As we keep growing with NEO Powered by Better and the Tinman AI platform overall, we anticipate handling more of these specialized categories, delivering higher gain on sale margins and greater efficiency thanks to our technology. Having demonstrated the Tinman AI platform through NEO Powered by Better while the whole mortgage industry looked on, we now find ourselves flooded with inquiries from other mortgage teams and companies eager to shift their operations onto the Tinman AI platform.

Vishal Garg

executive
#4

There's a huge opportunity ahead of us when it comes to other traditional mortgage originators, and we're making good headway working through a strong pipeline of prospective clients and partners. I feel great about what we've accomplished independently through both our D2C channel and our Tinman AI channel. And to push things even further, we're now working on connecting those 2. What really excites me is the testing we've got underway right now, where our AI, using a comprehensive set of parameters about the customer and the home they're purchasing, selectively pairs preapproved D2C purchase customers with localized NEO Powered by Better loan officers who know their specific geographic markets inside and out. These loan officers have also built wonderful ties with the realtor community in those markets going back decades, which means they can slot right into the workflows those realtors already like to use. Once this is running at scale, we anticipate the possibility of a meaningful lift in conversion plus added volume and revenue at very attractive margins for the D2C side of things as well as for NEO Powered by Better. We're still in beta mode, but we think this strategy can keep enhancing the unit economics of our D2C business down the road, since it brings in revenue from customers who otherwise wouldn't have converted on our online platform. Put differently, those preapproved leads would have arrived via D2C self-serve, figured out they'd rather have a high-touch experience with someone who knows their local market, and therefore been in danger of dropping off. Now the AI is evaluating, case by case, whether that lead would more likely have converted with a local market expert on the Tinman AI platform. And over the coming 6 to 12 months, our matching algorithm will keep learning and getting better — right in time for the 2026 purchase season. As a quick reminder, Better produces more than 250,000 preapproved mortgage customers annually. Applied to our average loan volume of close to $300,000, that points to a total volume potential of $75 billion. Yet our overall share of the market for purchase originations we actually fund sits below 50 basis points right now. Measured by home shoppers per year relative to the number of loans we preapprove, our market share is actually more than 10x that figure, at almost 5%; this highlights the enormous opportunity we have by unlocking another path to conversion for the customers who come to us seeking a preapproval. And turning back to our multipronged distribution approach, we serve the customer while also powering banks, credit unions, and other larger mortgage originators that want to license our Tinman AI software so they can become more efficient and customer-centric. We've built a highly fine-tuned platform for our own business and customers, and now we're seeing demand from others in the industry who want to license our software directly for use in their own operations. Many banks and credit unions are reexamining the mortgage business now that the regulatory climate has grown more accommodating. Even so, originating mortgages has mostly been a money-losing activity for banks, since their origination costs run high. That's the gap our Timman AI software fills. What Timman AI software really delivers is a mortgage-in-a-box solution: banks can deploy our software and, if they want, also tap underwriting and sales resources we make available. Once a bank or credit union adopts Timman AI software, there's no longer any need to sink money into numerous separate systems or to spend millions of dollars on integration, upfront expenses, implementation, seat licenses, or anything along those lines. Finished are the times when eight distinct software platforms had to communicate with one another, held together by middleware and by consultants demanding hundreds of thousands of dollars merely to connect those systems, all while lacking a single unified data set for any form of machine learning to run on; we're transforming this industry completely. On top of that, our pricing approach is radically different; it's what's increasingly gaining traction among AI firms under the name outcome as a service. Our customers are charged purely on the basis of funded loans, so their cost event for processing, fulfilling, underwriting, and even selling a loan is tied straight to the revenue event, removing the transaction risk for these partners — a huge point of distinction versus how software has traditionally been priced in mortgage and in financial services more broadly; we're glad that the first bank partner on the Tinman AI software platform has started funding loans through it. Through this partnership, we're driving their whole mortgage platform on the software side, from click to close, with their sales and operational staff, across every product category they provide, covering non-QM and other niche products, all entirely on Tinman. Bringing them live — not only for conforming, FHA, and VA but also for niche non-QM offerings — required slightly less than 3 months between LOI signing and loans moving through our software; and given our pipeline, we think this can scale to more than $4 million per month in monthly revenue over the near to medium term. We think a still bigger addressable market awaits within the mortgage ecosystem for a comprehensive one-stop software solution driven by Tinman, the industry's top AI engine. For context on the size of this opportunity: more than 5 million mortgages were originated on the Encompass platform during 2024. If we manage even 1% penetration of Encompass's customer base at our present pricing, we believe that would generate an additional 50,000 new loans and $75 million in revenue at software margins on a better per-year basis. Our partner pipeline is incredibly robust right now. These are companies looking to either break into mortgages or expand within the space, and they want to do it through the Tinman platform. What that gives them is the capacity to grow almost without limit on the sales and underwriting side, all without adding headcount — and that headcount problem is exactly what dragged down the whole mortgage sector during the downturn from 2022 to 2024. The last thing most mortgage chief executives want is to relive that experience. Fortunately, with Betsy they get virtually unlimited scalability, feature parity with a conventional loan officer handling refinances, plus the flexibility to ramp marketing spending up or pull it back depending on what their balance sheet requires. No one else in mortgages offers this. It stands alone in consumer lending too, and we're extending that capability to several of the biggest financial services firms across the United States. Once again, we look forward to updating you as these partnership conversations progress and we move into execution and launch with these partners.

Vishal Garg

executive
#5

To sum things up, our D2C operation has consistently led the way in stretching what technology is capable of achieving within mortgage lending, but at the same time, we're achieving significant progress toward greatly expanding how Tinman gets used—it stands roughly years ahead of everyone else in this sector—by diversifying in two directions: Tinman AI functioning as a platform serving other mortgage originators, plus Tinman AI operating as a software service addressing the mortgage industry's fractured tech stack. As we look toward the latter portion of 2025 and what comes after, the possibilities in front of us have never been more thrilling. Our attention stays fixed on strengthening our go-to-market approach, with growth serving as our North Star, together with ongoing expense management and channel diversification, all aimed at reaching profitability on an adjusted EBITDA basis within the coming 12 months. Although we'll keep investing in building the premier AI platform for mortgage lending, Tinman, so we can enhance our customer experience and push labor costs down further while making our platform more efficient and scalable, in the end, what we're aiming for now is driving this business to profitability, and we expect to get there on an adjusted EBITDA basis during the next 12 months. I'll hand things over now to Kevin Ryan, our Chief Financial Officer, who is going to go over the quarterly performance and our financial strategy moving forward. Kevin?

Kevin Ryan

executive
#6

Thanks, Vishal. Like we've talked about on earlier calls, although the market keeps presenting difficulties and macro volatility remains elevated across our sector, we're still achieving significant movement toward our objectives: growing volume and revenue while simultaneously keeping expenses in check and getting more efficient. Reaching profitability over the medium term has been the aim; at this point we possess both the route and the clarity needed to steer toward adjusted EBITDA breakeven by Q3 2026. The forces behind that are volume gains in our direct-to-consumer and Tinman's platform channels, per loan contribution margin that keeps getting better, further build-out of higher-margin channels such as Tinman AI platform and Tinman AI software, better pricing, plus ongoing cuts to corporate and vendor costs. Worth pointing out — these avenues for growth each carry different levels of expansion and profitability profiles, and they'll shift depending on how the broader macroeconomy moves. Consequently, getting to adjusted EBITDA breakeven probably won't follow a straight line quarter to quarter, and we don't expect an identical amount of burn reduction every single quarter. During 2025's second quarter, measured against the same period a year earlier, funded loan volume rose 25%, hitting roughly $1.2 billion, while revenue climbed 37% to $44 million, powered by funding a greater number of loans across both our D2C and Tinman AI platform channels. Our adjusted EBITDA loss came to about $27 million. Looking at channels, 64% of second-quarter funded loan volume originated via direct-to-consumer, while 36% came through Tinman AI platform, together with B2B. Breaking it down by product, funded loan volume consisted of 67% purchase, 20% second lien, and 13% refinance. Comparing quarter over quarter sequentially, Q2 funded loan volume rose roughly 39% and revenue increased approximately 36%; that revenue gain stemmed from higher volume out of NEO Powered by Better, which carries better gain on sale margins, our ongoing effort to push pricing higher, and a favorable tailwind from a loan loss reserve release. We intend to keep fueling growth via Tinman AI efficiencies, broadening our distribution channels, and refining marketing, all while offsetting these growth costs with additional reductions in corporate and fixed vendor expenses. Now onto expenses. In the quarter, total expenses fell about 3% in Q2 versus Q1. We're still developing our Timman AI platform and Timman AI software channels, embracing productivity-driven savings by deploying AI throughout the mortgage business and pushing costs lower still within our corporate functions. Using AI to move the business toward growth and profitability excites us, much like the strides we made between 2016 and 2021, when originations expanded by more than 100x; now, briefly, on our balance sheet and capital position. As we mentioned on our previous call, in April we completed a significant debt restructuring alongside our partners at SoftBank. The accounting entries are somewhat intricate and are detailed in the release. The broad strokes: we boosted our GAAP equity by more than $210 million, and we substantially cut our corporate debt. We closed out Q2 of 2025 holding $241 million across cash, restricted cash, short-term investments, and assets held for sale. On top of that, our ties with financing counterparties remain solid, with 3 warehouse facilities providing total capacity of $575 million as of June 30, 2025. One thing we're really enthusiastic about is that loan volume on the Tinman AI platform keeps expanding in line with forecast, and we anticipate more than $500 million in AI platform originations during Q3 — that's over 16% growth compared with Q2. Looking at the whole of 2025, funded loan volume should rise year-over-year, powered by tailwinds from growth initiatives such as the Tinman AI platform, though this will be partly offset by ongoing macro pressure and the loss of our Ally business, which represents a headwind of roughly $1 billion. Over in the U.K., we were delighted that Birmingham Bank increased its loan book 90% in Q2 on a sequential basis relative to Q1 of 2025. Even as we keep working to exit our noncore U.K. assets, we anticipate that divesting 3 smaller noncore U.K. businesses will begin benefiting our adjusted EBITDA in the latter half of 2025 thanks to those dispositions. We foresee additional gains in our adjusted EBITDA losses for 2025 relative to 2024 via a mix of AI-driven conversion rate improvements, efficiency gains, and ongoing corporate cost reductions; with that, I'll hand things back to the operator for Q&A.

Bose George

analyst
#7

So when you mention partners attempting to get into this area by leveraging your technology, could you describe what kind of companies they are? Are these finance outfits that haven't yet found an effective method for delivering the offering? Or whatever detail you can give would be helpful.

Vishal Garg

executive
#8

Absolutely. I'd say there are — within this area, allow me to lay out a few distinct categories of partners coming in. First, you've got the next-gen wealth management firms that are looking to keep climbing the ladder so they can go head-to-head with the UBS, Charles Schwabs of the world, right, and turn into full-scale wealth management platforms. The single wealth product — the single lending product that essentially every wealth management company truly requires is a mortgage. So we've teamed up with one of the top 3 kinds of robo-advisers investment firms and are running a beta test on a product; and we're receiving more and more inbound calls from organizations of that type who have — possess customer bases ranging anywhere from 1 million to 20 million customers who are interested in providing a mortgage or home equity product as a supplementary offering alongside their wealth management services. Second, you've got your conventional fintech lenders. Every one of the large fintech 1.0 lenders that have led the way in personal installment loans, have led the way in BNPL, now have substantial customer bases, once again, somewhere between 1 million to 10 million customers across these platforms. Those companies are now attempting to work out how to move into home equity. And the conventional home equity software offering or software platforms are essentially absent because the home equity market basically perished following the global financial crisis and has only revived over the last couple of years; and there are genuinely very few white labelable third-party origination platforms that operate seamlessly across the entire stack from click to close on home equity, and especially those that don't obligate you to send the loans specifically to one takeout, one securitization takeout. So there are 2 or 3 other platforms that license out their software, but they obligate you to send the loans to a specific takeout. They don't always permit you to keep the loans on your own books, all those kinds of things, or the fees they charge are quite exorbitant; so we're seeing a lot of influx on that front on the home equity side from those fintechs; and then third is simply the mega fintechs, the ones with the 20 million-plus customers, right, who essentially sat out the mortgage boom in 2020 and 2021, now observe rates declining and are beginning to see customers having demand; and many of those require a highly scalable solution. Many of them have brought AI into their own operations for their companies and want options capable of delivering essentially limitless scaling if the interest rate situation shifts; so those are conversations we're having too; with several, we're in beta; with several, partnership launches have already happened; and as those grow, we're eager to share more with you once they turn into a meaningful contributor to our revenue.

Bose George

analyst
#9

Then, when you mention, Kevin, that the cost to originate is at the industry average and coming down, how much volume is broadly needed on the overhead side just for that benefit to really flow through to the bottom line in a more meaningful way?

Kevin Ryan

executive
#10

I believe it already does. If you turn to Page 18 in our deck regarding unit economics, where we stand on D2C, correct? In my view, D2C is burdened by a high CAC and the manner in which that CAC is recognized, especially for purchase loans, correct, since CAC is kind of recognized in the period, yet the loans require 6 to 9 months to be made. Even so, for D2C in the second quarter of 2025, you'll notice the labor expense for each funded loan sits under $2,500, which compares favorably with what the industry typically spends — between $6,500 and $7,000 — to get a loan funded. Meanwhile, our per-fund data expense continues to shrink as our conversion rate climbs higher. Frequently, we gather consumer data, issue a preapproval, and acquire credit income asset data that we've paid for, only to see those consumers not convert. That's why improving the conversion rate matters — as we noted during the call, Betsy helped us lift it by roughly 30% this quarter, and as you can see, the data cost for each funded loan has dropped from $1,200 down to $800. We think that number can continue to come down, partly from scale efficiencies and partly from getting better at timing and executing how we retrieve the data customers need. On top of that, gain on sale revenue climbed almost 10% from one quarter to the next — once again, by responding more nimbly to customers and connecting with them on their terms, whenever they choose to reach out, we've been able to offer a narrower discount on our products and, on average, improve our gain on sale. That's worked in our favor. Regarding D2C specifically, if that were our sole operation, getting to roughly break-even would require growing it by two or three times, given our current contribution margin position and the monthly burn rate's downward trajectory. However, once you factor in the B2B platform side — the Tinman AI platform business — the margins there are considerably higher, nearly two to three times what the D2C business generates. What we have to do is expand those and get the volume up to levels comparable with what D2C delivers, and that's what gets us to the point where we're not losing money. That's the reason I'd say that after four years, this is the first time we can see clearly and with real confidence the path to breaking even.

Brendan McCarthy

analyst
#11

I'd like to begin with Betsy's lead-to-lock conversion rate. You said it went up significantly. What's actually behind that? And what customer interaction dynamics are driving that gain?

Vishal Garg

executive
#12

The main factor is that Betsy's capabilities have grown. Before, she only handled answering questions and collecting missing information. Now Betsy can guide someone the entire way to lock, and she'll even ask about processing and request any documentation that's missing. On the back end, for every loan the underwriters work on, Betsy is re-underwriting it, including cases where a loan might need to be suspended or denied after lock. As a result, the options available to consumers are much better, along with the smart workarounds involving the many investors in our marketplace. All of that is producing much greater efficiency. We're approving a bigger share of our customers. We're instantly finding solutions for them — for instance, take a consumer who comes in and whose debt-to-income ratio means they can't get preapproved for a mortgage or a refinance. Say they came in for a rate term refi, but there are 2 debts that, if they also pulled out a little cash, they could pay off, and then their DTI would qualify. Normally, it's quite hard for a human loan officer to run all those calculations; their training is in sales, not math; but Betsy can essentially do it instantly within our flow, and that's rescuing a pretty meaningful number of people and qualifying them in a way that other loan officers — and certainly other loan platforms — just don't. Much of what we're doing here is genuinely interesting, because we're learning from our retail loan officers about the things that make them more successful, letting them both command premium pricing in the market and effectively help more customers. And those very things our retail loan officers are teaching us, we're now building into our D2C workforce and, even more importantly, into Betsy, so Betsy learns that functionality and delivers it for the consumer instantly.

Brendan McCarthy

analyst
#13

Thanks for that explanation, Vishal. Now, shifting over to the B2B segment, and more specifically Tinman AI as a piece of software — could you give us some color on what that pipeline looks like? I recall you previously brought up a small to medium-sized bank that was in the pipeline and might have been closer to a potential partnership in the near term. Is there anything new to share on that?

Vishal Garg

executive
#14

Absolutely. Loans are being pushed through it right now. It's genuinely fantastic. In terms of mortgage operations, this bank is roughly $1 billion in size, and they've completely embraced it; we managed to bring them live on a brand-new mortgage software platform — something that normally requires 9 months plus millions of dollars spent on consultants and integration — and they were operational in under 90 days. As the volume flowing through their pipes increases, we believe that could turn into a fairly meaningful revenue stream for us. There are a few other major partners in the pipeline, including one of America's top 10 mortgage companies; we also have — many of the large fintechs I brought up earlier when answering a question on this call. So we're extremely excited about that software business and about enabling other mortgage companies, or those in the fintech space, to leverage the software effectively to originate loans in a manner that beats the traditional kind of 8 different systems stack that most people rely on today.

Rayna Kumar

analyst
#15

Abigail Rutter here, filling in for Rayna; the testing that you all referenced sounds like it's going quite well. Could you all give us any extra detail to flesh out this opportunity? And after that, what's your expectation for how it will perform across the cycle?

Vishal Garg

executive
#16

So, our expectation is that this becomes a software operation with margins far above even the platform operation, which runs at a 40% margin. I mean, $1 billion has gone into building this software. The largest competitors out there hold sizable or basically monopolistic positions and command really hefty chunks of market share, and they began as Fay disk-based software. Even the top competitor today permits just a single individual inside the loan file at any given moment. I mean, without judging it, what's the AI agent supposed to do? She'll tell the loan processor to check out of the file, as though it were a library book, so she can get in there and record an entry or alter a fact. It's just — my view is it's simply open season on that whole area. We've brought in — made a few recent additions to staff up that operation. And before long, you'll be hearing from us regarding several major contracts being signed and implemented. Once those are live and operating, we'll be in a position to discuss them. As for how big this could get, given the basic distinction in our approach, since we deliver one full-stack solution, customers will take it up, and then they'll take it up across 100% of what they produce. At first, they'll run it alongside whatever incumbent solutions are currently in place; they'll begin with home equity, HELOCs, after which they'll shift additional products across. And we're working through that whole process, mapping out the complete value chain along with the extractions. And yes, I think if you search for us on social media, if you search for us on something like LinkedIn, you'll spot some of the developments rolling out frequently in the software and with the partners we're deploying the software for.

Rayna Kumar

analyst
#17

Could you expand a bit further on the home equity operation and its volume? Also, any additional detail on what has been propelling, or is currently propelling, that expansion?

Vishal Garg

executive
#18

Of course. So comparing one year to the next, coming out of Q2 '24 — during Q2 '25 we originated $240 million in home equity volume, versus $90 million in Q2 '24. That works out to 260% growth. At this point, I believe we hold the title of America's fastest-growing lender in the home equity space. A lot of what powered our success in refinancing also applies here—the same qualities that let us go from $500 million in refi volume in 2016 all the way to $58 billion in 2021, spanning 5 years, which works out to a 100x increase. Those qualities come down to a direct relationship between us and the consumer, plus our ability to deliver what we do better than anyone else: a process that's more affordable, speedier, and simpler, run marketplace-style instead of feeding into a securitization shelf. That means we can serve several different investors and give consumers the best shot at approval, which in turn has allowed us to get the most out of our D2C margins and D2C dollars. Take an originator — call it originator X — whose lending standards are quite tight and who sells into a securitization shelf while being limited by what the rating agencies will allow. Those things are not easy for that originator to adjust. Now imagine they buy a click online, and there's some probability, X percent, that the click turns into a conversion. By the same logic, when a mortgage broker sends a customer over to originator X, the odds of that customer being approved are slim. But if that very same online click could be converted at a higher rate — a higher approval rate, and better, since a marketplace sits behind the whole thing — then our capacity to expand through the D2C channel becomes quite significant. To repeat, HELOC plus home equity origination has reached a $1 billion run rate, and this offering has only existed for under two years. We're expanding it incredibly quickly, correct? The figure stood at 148 during the prior quarter. That means quarter-over-quarter growth for this product exceeded 35%. So we simply believe the opportunity ahead is enormous, and many of the product's characteristics align with areas where we perform especially well.

Kevin Ryan

executive
#19

Right. I'd also point out that how well customers are taking to this product has risen significantly. It decelerated following the crisis, which Vishal brought up earlier. Even so, mortgage rates remain elevated, and the increase in home values since COVID has been extremely strong. That means the $1 billion figure could quite readily reach $2 billion. We'll simply continue pressing forward with this product, since we believe the consumer — it can genuinely assist the consumer during this portion of the cycle.

Vishal Garg

executive
#20

Furthermore, any HELOC — any home equity borrower we sign up at present — will eventually become a refinance once rates decline. That refinance requires no customer acquisition cost at all; and, as you're able to see, for us D2C customer acquisition cost amounts to between 45% and 50% of revenue; therefore, once we remove that expense, when rates fall again, we will move each of those HELOC borrowers into either a cash-out refinance or a rate refinance, and doing so will run us $0 in customer acquisition cost; so, once more, it's simply about bringing consumers onto the better platform and getting them ready for the refinances ahead. Your last question is from Eric Hagen of BTIG.

Eric Hagen

analyst
#21

Could you describe the way the software might tweak or customize things when rates move up and down? Is the AI capable of becoming savvy regarding conditions in the capital markets so it can stay in front of the market, so to speak? We're all aware, for instance, that mortgage banks may be sluggish when it comes to reacting to changes in rates, yet does the AI assist in boosting how quickly a consumer is presented with a rate, depending on what the market is doing?

Vishal Garg

executive
#22

Indeed, it renders it instantaneous. At this point it has begun generating the notifications that are sent out to consumers. It's allowing us to move into capabilities that will make the trading of interest rates, or the trading of mortgages, extremely close to the trading of equities. And it can simply bring about optimal execution and reprice our offerings at a remarkably quicker pace. Thus a conventional bank could take some time to refresh the rates shown on its website or via its distribution channels. Typically, those in mortgage banking receive daily rate sheets or best X. With the AI, there's no requirement for a rate sheet; across every permutation, spanning all 45 investors and thousands of products, the AI can instantly locate the optimal match for the consumer, suit their requirements, and deliver that array of choices to them. It's genuinely transformative. In my view it is hugely undervalued in this market, where the cost of credit is constantly shifting; and it will — you're going to witness something like a Robinhood moment in this industry, just as you witnessed it with equities.

Eric Hagen

analyst
#23

Yeah, that's really interesting. A follow-up on the conversion rate. Could you share more about how the Tinman platform helps keep borrowers engaged and moves them from lead generation to the next stage? We get that the platform is great at lead gen, but how does it systematically drive a higher conversion rate, especially given the many different entities using the platform?

Vishal Garg

executive
#24

Sure; historically, the Tinman platform has done very well serving consumers who have great credit, make good money, and are comfortable making big financial decisions online, right? So even though the TAM in mortgages is huge, the TAM for consumers willing to handle these things entirely online and on their own has always been smaller. So what we've kept trying to do over this period is that whenever live human help is needed, staffing for it turns out to be really expensive, right? For instance, someone might say they drop their application at some point in the process, right? We could build — we have built triggers that allow us to spot when someone has dropped their application. But because all the labor in the industry, like loan officers and all sales staff, has to be licensed, and those licenses differ by state, unless you have tens of thousands of people sitting around waiting for these triggers to fire and paying them through payroll, it's very hard with our low-cost D2C model $148 million to reach the consumer at the moment of decision or to reach the consumer when they're unsure, fatigued, or confused. And now, we can do that instantly and have Betsy make an outbound call. As you can see, interactions involving Betsy have grown enormously over the last 3 or 4 months, because it has just kept improving and become more and more integrated; it's almost natural now between Betsy and Tinman. So it's kind of becoming one system, with Tinman as the computation and matching engine, while Betsy supports the human — yet in many cases, like replacing the human, when the human just isn't available. And I think that has truly let us keep growing. For example, we raised unit volume by 35% in the prior quarter, while we cut headcount on the sales team.

Vishal Garg

executive
#25

We appreciate everyone sticking with us while we work to create the top AI mortgage platform in America, and by doing that, assist customers in obtaining a superior mortgage, a superior rate, and a superior process, which helps them enhance their financial security and enjoy an improved life for themselves and their loved ones. The last 5 years have proven very difficult for our company because of market conditions and interest rates, but at this point we're once more attacking aggressively, vigorously chasing expansion, generating revenue from the platform we've constructed and the AI we've created upon it, and accomplishing this regardless of wider economic and mortgage market circumstances. We currently possess a distinct understanding of how to succeed under present conditions through our progress with our Tinman AI platform and our Tinman AI software business plus enhancements in unit economics, all of which combine to let us state plainly today that we will reach the point of breaking even as a company on an adjusted EBITDA basis within the next 12 months. That, I believe, is the key point all of you ought to carry away from this call; we have a route out of an exceptionally awful situation now, and we have virtually limitless capacity should the interest rate landscape shift. We're grateful for all your backing. We're grateful you've remained with us, and we anticipate bringing you much more positive updates throughout the year to come.

Operator

operator
#26

This concludes today's conference. You may disconnect.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Better Home & Finance Holding Company transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.

Get the API View API docs →

This call discussed

For developers and AI pipelines

Programmatic access to Better Home & Finance Holding Company earnings transcripts and 132,000+ others is available through the EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments, full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.

More tools for stock research

OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Better Home & Finance Holding Company or any ticker on your watchlist.

StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.