Prepared remarks
Hello everyone. Thank you for joining us and welcome to Blend's Financial Results Conference Call for the second quarter of 2026. I will now hand the conference over to management for their prepared remarks. Please go ahead.
Good afternoon and welcome to Blend's Financial Results Conference Call for the second quarter of 2026. I'm Meg Nunnally, Blend's Head of Investor Relations. Joining me today is Nima Ghamsari, our Co-founder and Head of Blend, and Jason Ream, our Head of Finance and Administration. Before we start today's call, I'd like to note that we will refer to certain non-GAAP measures which are reconciled to GAAP measures in today's earnings release and in the appendix of our supplemental slides. Non-GAAP measures are not intended to be a substitute for GAAP results. Unless otherwise stated, all financial measures we'll discuss today, including our profitability, refer to non-GAAP. Also, certain statements made during today's conference call regarding Blend and its operations, in particular our guidance for the third and fourth quarter of 2026, other commentary regarding 2026, and our expectations about markets, our strategic investments, product development plans, and operational targets, may be considered forward-looking statements under federal securities law. We caution you that forward-looking statements involve substantial risks and uncertainties and a number of factors, many of which are beyond the company's control, could cause actual results, events, or circumstances to differ materially from those described in these statements. Please see the risk factors we've identified in our most recent 10-Qs, our 10-K for the fiscal year 2025, and other SEC filings. We are not undertaking any commitment to update these statements if conditions change, except as required by law. The financial information presented on this call is based on continuing operations and any prior periods have been recast to exclude operations that are now discontinued. Lastly, we will be providing a copy of our prepared remarks on our website by the conclusion of today's call and an audio replay will also be available soon after the call. I'll now turn the call over to Nima.
Thanks Meg, and welcome everyone. The second quarter was another disciplined, profitable quarter for Blend. Revenue came in near the high end of our guidance range and non-GAAP operating income came in above the high end, and Jason will take you through all the financial details in a few minutes. But today, I want to spend my time on the two pillars of our strategy. Autopilot, the agents we build for our customers, and Blend 3.0, the agents we are building inside Blend to help us do our work to serve our customers faster, better, and cheaper. Let me start with Autopilot. The big news for us last quarter was that Autopilot became commercially available on July 1. If you're following along on the webcast, I'd invite you to advance the slide with the words, our AI strategy, guiding our customers at the top. This is the first of two slides that I'll reference today, and the charts on these slides are also available in the supplemental slides on our Investor Relations website. The chart shows cumulative loans processed by Autopilot from a standing start in February to more than 45,000 today, and it's still compounding. The curve is the proof behind everything I'm about to tell you. During this four-month window, more than 65 lenders activated Autopilot. They stress tested, surfaced the hard edge cases, and shaped what we shipped. And the preview data backs up why this matters. We're starting to see evidence Autopilot is driving faster clearance times, higher conversion rates and potentially reducing fulfillment costs. Based on our preliminary data of these loans that have gone through our system, our customers are seeing a 10% to 15% improvement in pull-through rates and 2 to 4 days of cycle time improvement. Furthermore, we estimate that Autopilot is automating 4.5 hours of loan fulfillment tasks on average per loan. This is huge and our customers are just beginning to grasp the potential. Now that we are commercial as of July 1, 6 lenders have already signed contracts that include Autopilot, including Onity, which is one of the largest mortgage servicers in the nation, which is also building its own experiences on top of Autopilot through our Autopilot MCP server. On monetization, we're executing the plan we described in May, with customers signing flat-fee one-year contracts for full access. There's going to be a base level of intelligence built into our workflows, but the paid tiers are where the full product lives. We call it underwriting intelligence, where Autopilot is reading documents, running calculations, reconciling against guidelines, and driving the full loan file forward. Over time, our intent remains to move the paid tiers to a per funded loan model, just like the rest of our Mortgage Suite. And when seat-based pricing, which we don't think survives the agentic world, when software does the work, you can't charge by the person. It doesn't make any sense. And others are also moving to consumption models to get paid for the activity their AI generates. We made a different choice, which is we get paid on success. We get paid on the outcome, not for the tasks along the way, but for the outcome. So if Autopilot does 10x the work on a file that never closes, our customers shouldn't pay 10x more for it and under our model, they won't. Our revenue scales with our customer success and that alignment is the business model we've always had. Agentic AI just makes it more valuable and more scalable. Since I get asked about the competitive landscape constantly, let me be direct about it. What gives Blend the right to win? Our answer comes down to four advantages that are hard to replicate. First, where we sit. Blend is the borrower's first point of contact. And because we're there, we see the problems and we fix the problems as they come along. More than half of borrowers apply outside of business hours, and over 90% of people who complete an application do so within 24 hours of starting. Autopilot catches friction the moment it happens, at 10:00 p.m. on a Saturday, not Monday morning, the exact moment the borrower intent is the highest and the engagement is the highest. Next, our data. The question behind every AI question I get is, in a world where anybody can call a frontier model, like Claude, is a software like ours replaceable? We believe the answer is no, and the reason is our data. Models are converging and everyone has access to the same models, including us. But what isn't a commodity is what it takes to get the model to perform on a mortgage that has tons of context, tons of loan documents, guidelines, and lots of things that have to be taken into account to make the right next action on the loan. And that's a combination of 15 years of experience and tens of millions of loan application data processing through our platform, which you can't buy. And you can't synthesize 15 years of real borrower behavior. And that compounds. I actually see this in our early Autopilot benchmarks against a typical Claude plus skills. And we see that Autopilot performs better in those tests, and much better and much cheaper, about one-third the cost. Third, the harness. Autopilot is not a wrapper around a generic model. It runs inside of the infrastructure. The data layer, the integrations, the compliance architecture, orchestrating what the agent sees, what tools it can use, which guidelines to apply to a specific loan in front of it, and what happens when it isn't sure. It hands the file back to the loan team. People stay in control of the decisions that matter, and this harness is part of what makes Autopilot more accurate and cheaper than a generic harness. And last, our relationships. We've spent 15 years building alongside lenders, and Autopilot was built the same way, with lenders, for lenders. And with Autopilot MCP, we've opened up our infrastructure so customers and other technology providers can build on top of us rather than around us. And when you step back, this is the bigger thesis. Blend is and can be the agentic infrastructure for relationship banking. The original promise of banking was a relationship, a lender who got to know you and could make a call based on more than just a credit score. That promise didn't disappear because bankers stopped carrying it. It disappeared because there's so much process, and there's so much manual work and manual effort that has to go in to every single loan. But when Autopilot is handling that grunt work on the loans and helping the people who historically did that focus on the customer, when Autopilot is handling those conditions, the follow-ups, the questions that come in at 2:00 in the morning, and the loan officer can get back to the capacity of serving the customer. I started Blend with my co-founders back in 2012 with a simple thesis that the mortgage process should drive itself. Not because humans aren't needed, because the right technology can handle everything that does not require a human. And Autopilot is finally that thesis arriving. And because of what we're seeing in mortgage, the most complex, most heavily regulated process in consumer finance, we believe the same infrastructure extends naturally to home equity, deposits, auto, card, personal lending. And I think probably broadly, given how much we've honed the harness and the evaluations around Autopilot, beyond that to other aspects of underwriting. Agentic AI doesn't replace and doesn't need to replace relationship banking, but it makes it possible to have our customers, the lenders, be even more focused on their customers once again at scale. Shifting gears, if you're following along the webcast, I'd ask you to advance this slide with the words, our AI strategy transforming how we work at the top. This slide gives you a glimpse of how Blend itself is transforming and how it is going to look in the future. And we're calling that Blend 3.0. Where if you think of Blend 1.0 as the first 10 years where we built a market leading company and we're growing quite a bit, growing our market share, growing our customer base and rolling out our first product. And Blend 2.0 is the last four years where we were creating a profitable long-term sustainable entity. Blend 3.0 is an agentic first company. And that doesn't mean just for our products, but that also means how we work internally. And last quarter I described this pattern. An agent will, in the end state of Blend, take the first pass of the incoming work, especially the grunt work that I described earlier, before a team member even touches it and applies their judgment. So I won't repeat all that today, but I want you to look at the graph. And the graph shows one specific team in the company, which is actually one of our biggest teams, which is the engineering team. And if you look at the gray line, and you look at pull requests, which is the gray line in 2025, it was pretty flat throughout the year and during the holidays it trailed off. And then you look at the blue line this year, we've 3.6x our throughput as an engineering team with roughly the same headcount just since January of this year. 3.6x the throughput. I don't know if you remember, but I told you in May that we were running roughly 1.5x compared to what we were doing in January. And that was the average. That was just three months ago. And now we're 3.6x. It's a profound thing to think about, which means if we continue at this pace, we could be doing 10x as much throughput on the engineering team by the end of this year as what we did the end of last year. And I think that will continue to compound as an advantage for us to turn into velocity of fixing customer bugs, handling customer feature and enhancement requests, building the new things like Autopilot that power the future of our industry at a faster and better pace. And that's really important because the essence of any software company is, how do you serve your customers and create value for your customers? And I think that trend will continue to steepen as we adopt further AI within our organization, and as the models get better over time. But given the success that we're seeing in engineering, the next phase that we're doing, and what I believe is our job, is to take this to the whole company. And so this past quarter, we expanded our efforts in agentic first approaches, and we expanded to our go-to-market organization, where agents now are reviewing support tickets that come in immediately. And if it's a bug that needs fixing, it can open a pull request. If it's a simple response that they want or a configuration change, it can draft that up for a human to go and click, yes, that's right or no, that's wrong, I need to change that. And it's also doing work on our customer calls and helping draft coaching notes and follow-ups that used to take the teams hours and in our finance organization, where agents are doing the first pass of work on parts of our closing process. In every case, the process is the same. The agent takes the first draft. A person reviews and approves. And in some cases over time, I think where it's less of a security issue or less of a code issue or less of something that doesn't even need a human approval, I think eventually over time, we'll even not need human approvals for some of those things. It's different functions that I'm talking about, but it's one operating model of how I think the future of agentic technology is. And it's still early days, but this is one of the most active topics inside Blend right now. Our leadership team is meeting regularly about it because becoming an agent-first company is an operating decision, not a side project. It's a whole company effort to figure out how would we reimagine this amazing company and customer base from the ground up because we now get that opportunity because we're profitable. We have our house in order and all these technologies are accelerating right in front of our eyes. And I said in May that we aim to be in the top 1% of all companies in agentic AI adoption, and that's still the goal. We're tracking our progress by function. We're taking it one piece at a time, and I expect to share more with you in the coming quarters as this rollout matures. And lastly, I want to talk about growth, because I know that's a question on everyone's mind. Well, to start with the bad news, the macro is not helping us right now. Mortgage rates moved from 6.4% or so in May to 6.8% in recent weeks because of wars and things that are going on outside of our control. And that keeps activity in the market, especially refinance activity, but also purchase activity muted. And so I've always said we can't control the macro. I just want you to be aware of it because we're obviously paying close attention to it and it might sort of mute future quarters like Q4 if we expected a certain amount of refinance activity, and it doesn't come because rates are high, just more being aware of it. But that's okay. Our ultimate goal has been and is to generate long-term sustainable growth regardless of the macro environment. In the quarter, we signed 14 new deals and expansions, and there's two that I want to highlight. The first is a new logo with a large credit union that included Autopilot right out of the gate. So I think this will be the new norm with our customers as they sign with us. We have some more deals in pipeline that have the same shape, but it's the first time a customer chose Blend and chose Autopilot as part of that initial package. Agentic AI is becoming table stakes and our customers want it and need it in fact to be the best versions of themselves. And I think, like I said, it's a preview of how many of our deals are going to look going forward. The second deal I want to talk about is a cross-sell of our Rapid Refi and Rapid Home Equity into a top 5 credit union, which they signed with us a couple of years ago. They got rolled out. They're happy. And this is a great example of how we can expand and deepen with customers as we drive the initial projects to success. And our pipeline continues to build. On our last call, I told you our overall pipeline was up more than 40% year-over-year. And that overall pipeline is still growing. What I'm more excited about as the year goes on is a narrower one. We brought in new sales leadership at the start of this year, and our late stage pipeline, the deals we aim to close within the next quarter, grew nearly 40% just between March and June. Late-stage pipeline is what actually is near signing, so it comes with a lot higher level of visibility and confidence. Just to put some color around this pipeline, that includes another large mortgage customer, a top 20 financial institution, and a solid set of Rapid and Autopilot deals, which, for example, we expect two additional large Rapid deals to close in the coming months. Now, I want to be honest about timing. We expect our pipeline to become signed deals in the coming quarter or so, and maybe some slip. And those signed deals become revenue. But this is all dealing with some of the largest financial institutions in the country. So as they sign and they turn on, we expect them to show up in our financials in the medium term, but it does take some time for very large financial institutions to get through their governance around things like agentic tools. So with all that said, I want to say the direction is very clear. Our customers are excited about it. Our largest customers are really leaning in and my confidence is very high. Our sales discipline that Matt has put in place isn't just limited to new business. It extends to how we manage renewals. We're doing a better job of getting our customers discussions around renewals with us earlier in the process, which means more relationship aspects to working with them, renewing customers for longer terms, broadening the relationship at renewal and making sure that our pricing reflects the value we deliver, which Jason will give you some more color later on how we're thinking about this, where we're seeing renewals and upsells, where we're seeing churn. So let me give you that other side of the coin right now, which is the core of our customer base is renewing for longer and for more. So to wrap it up, the short-term market with the macro, it sort of is what it is, but we're staying disciplined and profitable inside of it. The medium and long-term is what we're focused on, 2027 and beyond. As we get Autopilot going commercial and growing that, the pipeline we're building, the speed we're gaining with the agentic transformation internally, I am extremely energized about what Blend looks like on the other side of this cycle. Not just for us, but for our customers. And ultimately, because we have a value-based pricing model, what that means for our financials and for our investors. And so with that, I'm going to turn it over to Jason to walk through the financials.
Thanks, Nima, and thank you to everyone else joining us on the call. We delivered a solid second quarter with total revenue of $33.8 million, up 7% year-over-year and near the high end of our $32 million to $34 million guidance range. Mortgage Suite revenue was $19.2 million, up 7% year-over-year, within the growth range we discussed on our last call. Funded mortgage loans on our platform were approximately 233,000 in Q2, up 14% year-over-year and in line with our expectations coming into the quarter. That volume growth was partially offset by a lower year-over-year economic value per funded loan, which came in at $79, consistent with the $79 to $80 range we guided to in May. As a reminder, the step down from $83 in Q1 is primarily mathematical. Q1 is seasonally the high watermark given its lower mortgage volumes, and higher volumes in Q2 mechanically lower the per loan calculation given that there are some fixed fee arrangements within our customer base. Consumer Banking suite revenue for the second quarter was $12.2 million, up 6% year-over-year, slightly above the high end of the growth range we shared on our last call. And professional services revenue for the second quarter was $2.4 million, consistent with our expectations. Turning to profitability, non-GAAP gross profit was $26.5 million, and our non-GAAP gross margin was 78.3%, up from 76.1% in the second quarter of 2025, and consistent with the normalized gross margin framework of 77% to 78% we described last quarter after backing out the one-time benefits we saw in Q1. That gross margin improvement is despite the model costs for Autopilot, which were relatively low but growing in Q2. We expect those costs to remain fairly insignificant relative to our P&L for the near future, but we will update you if we see anything different on the horizon. Non-GAAP operating expenses were $19.5 million in Q2, roughly flat year-over-year. Non-GAAP operating income was $7 million above the high end of our $5.5 million to $6.5 million guidance range and represented a non-GAAP operating margin of 20.6%, an improvement of nearly 6 points compared to the second quarter of 2025. Free cash flow for the quarter was $6.9 million. We ended the quarter with $44.9 million in cash, cash equivalents, and marketable securities, and still with zero debt. During the second quarter, we repurchased 11 million shares at an average price of $1.65 per share. Year-to-date, we have repurchased 22.2 million shares for $36.8 million, leaving approximately $13.2 million remaining under our most recent share repurchase authorization of $50 million. We continue to believe that share repurchases are an excellent use of capital, especially at current valuation levels, though future repurchases will also be balanced against our aim to maintain ample liquidity to run the business. Before I give you the specific numbers, I want to frame how we're thinking about the environment because it shapes everything that follows. On our May call, and consistent with what we've said historically, we noted that our own outlook for the back half of 2026 was closely aligned with Fannie Mae, which at the time was forecasting roughly 19% full year growth in mortgage market volume. Since then, Fannie has lowered that outlook to about 17%. Our current view is slightly more conservative than Fannie Mae, as we expect refinance volumes to remain muted in a higher for longer rate environment. Specifically, we struggle with Fannie's forecast showing refi growth in the back half of the year, despite higher interest rates. Fannie may update their forecasts, but until then, we're taking the more conservative view. Additionally, we saw an uptick in churn notices this quarter relative to recent quarters. These are customers who are notifying us that they plan to roll off of Blend, though the timeline and ultimate impact is still to be determined. In most of these cases, customers plan to move to lower cost or free point solutions. We expect the revenue impact to be manageable in the low single digits of annual revenue, but it's a dynamic we're watching closely and as such, we have reflected some caution in how we're thinking about the second half of the year. With that context, let me walk you through how we see the next two quarters. Starting with the third quarter, we expect total revenue to be between $31.5 million and $33.5 million, representing approximately negative 4% to positive 2% year-over-year growth. Underneath these headline numbers, we expect a total mortgage market of 1.2 million to 1.260 million units, which is up 5% at the midpoint, and Blend funded loan volume of approximately 200,000 to 210,000 loans, which is up 2% at the midpoint. This translates to Mortgage Suite revenue change of approximately negative 4% to positive 3% year-over-year. We expect economic value per funded loan of approximately $80 to $81. We expect year-over-year Consumer Banking suite revenue growth up between negative 5% and positive 1% in the third quarter, consistent with the moderation we discussed on our last call. And we expect Q3 non-GAAP operating income to be between $3.5 million and $4.5 million, implying a non-GAAP operating margin at the midpoint of approximately 12%. As a reminder, the third quarter expense includes our annual Blend Forum customer event, consistent with prior years, which we expect to drive approximately $1.5 million sequential step-up in sales and marketing expense quarter-over-quarter. Looking beyond Q3, I want to give you our current view on fourth quarter volumes. Fannie Mae's most recent forecast is calling for roughly flat market volume in the fourth quarter. As I mentioned earlier, that forecast may get updated at some point, but for now, our own outlook is a little bit more conservative, and we expect total market size of 1.105 million to 1.165 million units, or down about 11% year-over-year at the midpoint. Against that backdrop and factoring in the final tail on the roll off of the large customer we've discussed on prior calls. We'd expect Blend's fourth quarter funded loan volume to be approximately 180,000 to 190,000 loans, down approximately 10% to 15% year-over-year. While this is our current best estimate, I should note that the macro backdrop remains highly fluid and sensitive to rates. Finally, on Autopilot, as you heard from Nima, we are very excited about the early commercial momentum, but we continue to encourage you to be cautious about incorporating Autopilot revenue into your models at this juncture. We plan to provide additional information on the potential impact to our model as customers and prospects move through the funnel and we have more time under our belt. In summary, we delivered the quarter we aimed for. Revenue near the high end, profitability above the high end, and a strategic return of capital through our share buyback. The macro remains a headwind, and we've tried to give you a clear view of how we see volumes in the market. But the underlying drivers we control, our customer wins, our product velocity, and now the commercialization of Autopilot, are all moving in the right direction. And they set us up to reaccelerate growth as we head into 2027. And with that, let's open up the call to your questions.
Questions and answers
Your first question comes from Aaron Kimson with Citizens.
Nima, last quarter you commented that you thought agents and Autopilot provided a path to potentially see 10% to 15% incremental revenue growth in '27. Given you're now a month into selling the one-year flat rate Autopilot contracts, you've got 6 lenders signed, and the commentary on it takes a little while to get through governance at large financial institutions, does that sentiment hold today?
Yes, I don't want to give any additional further guidance there, but yes, the pipeline is both good in terms of what we signed, but also I would say, a large number of our largest—what surprised me last quarter and this quarter was how much our largest financial institutions were leaning in. And I'll give one anecdote about that, which is, some of the biggest banks in the country, it's always been an executive priority, but the way we're seeing the business line and the tech lines of these places push the rest of their organization to get this into place has been surprising to me. It takes a lot to go and say, we really need to make this happen for our business, but they're doing that. I think that's why our largest customers are leaning in and some are in either late-stage trials with us or in final approvals with their internal committees. But like I said, those things take time and we're very excited about it. Probably more importantly, and I said this quickly in the prepared remarks, the early numbers from Autopilot are very encouraging. We have about 50,000 loans that have gone through in the last few months. We get to see some of the differences between the numbers before and after for those lenders. It's really interesting to see, especially given how early it is and the dual tailwind we have of our own work around the harness accelerating, and the models that are underlying it getting better. The way that we've set this up is to become this compounding machine for us. It's early, but the early signs are more encouraging than I would have thought if you had told me in January of this year, we're going to be creating this product from scratch and here's where we are six months later.
Understood. Yes, I think that chart on Slide 8 with the cumulative loans processed every 2 weeks is fairly compelling. And then the second question I have, for Blend 3.0, how do you think about the balance between faster innovation and better products versus OpEx leverage? That's one of the main questions I get from investors, I guess, based on today's update, when does the 3.6x increase in pull requests turn into either revenue or operating leverage?
That's a great question. I think some of the revenue side is answered by the previous answer I gave, which is these products take some time to become a reality in a market as regulated as this. It's part of our moat actually, in that we have these great relationships with our customers that bet on us and give us right of first refusal on their hardest problems, which in this case is their operations of how do they underwrite and process mortgages better, faster, and cheaper. But it does take time. That's why I kept focusing internally on the medium term, which is 2027. As these deals mature, as they get live and they get at scale, how do we use that as a weapon for us, but not focus on how do I drive immediate results at the expense of maybe 2027 and 2028 results? That's the beauty of being a profitable company right now in this challenging mortgage market with the macro working against us is that we can focus on the medium term and long term, which I know our investors care about. What does this company look like with an agentic core product that's driving a lot of the growth and with a macro that comes back to life? So we're very focused on that. I'm proud of the way the team is executing in this moment. Obviously, there's a lot more work to do for us. And the level of focus and intensity that the team has put in place around these efforts has been great to see as someone who's been here from day one.
Your next question comes from the line of Pallav Saini with Canaccord Genuity. The medium term and long term, which I know our investors care about. What does this company look like with an agentic core product that's driving a lot of the growth and with a macro that comes back to life? So we're very focused on that. I'm proud of the way the team is executing in this moment. Obviously, there's a lot more work to do for us. And the level of focus and intensity that the team has put in place around these efforts has been great to see as someone who's been here from day one.
First off, Nima, it's great to see the impact Autopilot is having on the mortgage side. I'm just wondering, is Autopilot also available to your consumer banking clients, and maybe you can speak to the opportunity there if there are any updates and I'll follow up.
Yes, it's built into our platform. We have the baselines in place to support all of our customers, which they're excited about. It's a big request we get from customers about how can we help some of them have multiple product lines with us? How can you help another business line that's still manually verifying IDs or forms that get filled out and sent to them? We do not have that live yet, but stay tuned. It's something we're thinking hard about and we have some early customer engagement around it. To share a bit about our approach on building products like this: pick a few lighthouse customers, drive it deep with them, prove the value, and then expand it to the market shortly thereafter. And I don't mean over the course of a year; I mean over the course of a quarter or two and drive that as quickly as possible.
Got it. And you mentioned 2 deals in the credit union channel. Maybe we can get an update on the channel in general and what the pipeline there looks like.
We don't have a specific breakout of credit union versus bank versus independent mortgage company or other financial institutions. But we have a strong base of customers from the largest credit unions down to medium and community credit unions. They're very member self-serve first and aligned to our product principles of how do we help someone who's applying at 2:00 in the morning get a full approval. The Rapid suite is a great example of this alongside Autopilot, particularly how we help someone get a full approval and offer initially, and then a full approval in the middle of the night. That alignment with serving members makes us well suited for this segment. They want to serve as many members as they can, they love the self-serve member experience, and it fits well with our product vision long-term.
Your next question comes from the line of Ryan Tomasello with KBW. Your line is now open. Please go ahead.
This is Huan Chong on for Ryan. I just want to double click on Jason's comment on the uptick on churn notices in the quarter. I understand that the impact is pretty small, but could you share a bit more color here? Are these going to be like the AI native point solutions that have just gone on the market or maybe customers are moving to build in-house?
No, it's a good question. Really, we saw people going to existing low-cost providers. It's hard to call it a trend; we don't have enough data points yet, but one of the things I would say is look, we're suffering a little bit with this rate environment and the macro and what that's doing to mortgage volumes, but our customers are suffering at least as much. They're looking for ways to save money and it's a tough environment for them as well. So I think that's what we're seeing in the few cases that we've had.
Yes, that's helpful. I have a follow up. Just one of the broader trends across the mortgage market has been market share shifting towards the broker channel. How do you think this could affect Blend and where do you think trends are going?
We've noticed that over the past few years, actually, not a short-term trend. It went from maybe single-digit percentages to around mid-teens or high teens percent in the broker channel. But the thing I would say is our customers—whether IMBs of all sizes, banks, or credit unions—are leaning into this market. Many of our customers are betting on the future and betting that as they get these new technologies in place and make the most of our system and other systems they have in place, they will keep growing. This year especially, people were getting their house in order in 2024 and 2025, and this year we're seeing companies in our customer base decide it's time to invest in the future. That was probably part of the cause of some shifts to other channels where some companies couldn't invest in 2023 or 2024. But as the market stabilizes, our customers are investing in the future. For our customer base, people are leaning in. They want new solutions, better service for their customers, and improved internal operations. Given how focused they are on improvement, I think our customers will continue to win in this market.
Our next question comes from the line of Dylan Becker with William Blair.
Nima, for you on Autopilot, encouraging early interest, but wondering, and you gave some good statistics around the improvement in pull-through reduction in cycle time, and each of those on an individual basis is compelling, but wondering how you're thinking about your opportunity to orchestrate the entire funnel of the workflow. And all of those kind of individual points compounding on each other as it pertains to customer ROI. Maybe how that's driving conviction in the Autopilot strategy and really what's differentiated from some of those potentially alternative tools or capabilities here.
Maybe just to give some color, because I think this is a really profound question, let me explain how typical solutions in the market tend to work. A lot of what I see is new things that layer intelligence on top of existing processes—take a document that was uploaded and add an entry that asks, did this document meet the guidelines? That approach assumes the existing process needs more intelligence built in, which is the opposite of the background worker agency or ambient intelligence concept I've described. It's essentially the antithesis of Autopilot. You're spot on in asking how we orchestrate the entire process. A lot of our vision for Autopilot and particularly coming together with Rapid is that Rapid gets you a real-time offer and then Autopilot should get you from that real-time offer to clear-to-close in the first session. That's where the market needs to go: a consumer can come through, get a self-driving process where they get a great personalized offer and they're ready to close, all at 2:00 in the morning. That's possible today regardless of legacy infrastructure because Autopilot is an open harness that can hook into our system or other systems. It focuses on the work that needs to get done. It's not there yet in every case to get a customer from offer to clear-to-close in one session, but that's the direction. We're working toward loans ready to close at 2:00 a.m., with the system checking documents, requesting new information, and working with the consumer in real time to get that information. That's the direction intelligence is heading. There's no reason in a market like this that consumers should wait 48 hours or two weeks to know if their application is approved when the technology can enable much faster outcomes. Having worked in this market for 14 years, I understand the depth and complexity of thousands of pages of guidelines and hundreds of pages of loan documents. The Autopilot harness is set up to handle that context in an incremental underwriting assistant fashion. That's where the compounding benefits will come into play for customers who will be able to do things at rates they didn't think were possible.
Got it. Very helpful. Thank you, Nima. And maybe a quick one for Jason. Appreciate the color on the mechanics of the outlook, if we look at EVPFL, that's brought into Q3 expected to kind of step up, I guess, just clarity around how much of that is kind of the mechanics of lower volume uplift versus maybe incremental product cross-sell. And I know we're not saying Autopilot is layering in any capacity, but if at all any extent that's driving some of that uplift as well.
Our volume expectations aren't that much different for Q2 and Q3, so I think it's mostly just the dynamics in the business. We're not guiding to a very different number; it's not a big change either way. I would think about Q2 and Q3 as sort of steady state with each other.
There are no further questions at this time. This concludes today's call. Thank you for attending. You may now disconnect.