Home / Transcripts / Blend Labs, Inc. (BLND) · May 7, 2026

Blend Labs, Inc. (BLND) Earnings Call Transcript

May 7, 2026

New York Stock Exchange US Information Technology Software earnings 49 min

Earnings Call Speaker Segments

Meg Nunnally executive
#1

Thanks for joining us this afternoon for Blend's conference call covering financial results from Q1 2026. Meg Nunnally here — I lead Investor Relations at Blend. With me are Nima Ghamsari, who co-founded the company and heads up Blend, plus Jason Ream, who oversees Finance and Administration. Ahead of getting into things, a quick note: some of what we discuss involves non-GAAP metrics, and you can find reconciliations to GAAP in the earnings release we put out today as well as in the appendix of our supplemental slides. These non-GAAP metrics shouldn't be viewed as replacements for GAAP results. Except where we say otherwise, everything financial we cover on this call — profitability included — is on a non-GAAP basis. On top of that, some remarks made on this conference call concerning Blend and how it operates — specifically our guidance for Q2 of 2026, additional commentary about 2026, and what we expect regarding markets, strategic investments, product development plans and operational targets — could count as forward-looking statements under federal securities laws. A word of caution: forward-looking statements carry substantial risks and uncertainties, and various factors, a great many of which lie outside the company's control, might lead actual results, events or circumstances to diverge materially from what those statements describe. Take a look at the Risk Factors laid out in our latest 10-K for fiscal year 2025 along with our other SEC filings; we're not making any commitment to update these statements should conditions shift, unless the law requires it. Financial information shared on this call reflects continuing operations, and earlier periods were recast so that operations now classified as discontinued are excluded. One final item: our prepared remarks will be posted to our website by the time today's call wraps up, and an audio replay should be ready shortly afterward too. Over to you, Nima.

Nima Ghamsari executive
#2

Thank you, Meg, and hello to everybody; the past 2 months since our previous call have been a real whirlwind; our Q1 results came out today, and Jason will dig into them, but revenue and non-GAAP operating income both landed above what was anticipated; during the quarter we also closed 15 new deals and expansions, among them an eClose agreement with a top-20 bank, plus a new mortgage deal with a different top-100 bank. As of March 31st, our pipeline has grown over 40% compared with a year earlier. That figure leaves out the Autopilot pipeline, which I'll get to shortly; however, during those 2 months the world has changed beneath our feet, with more global conflict, inflation and mortgage rates moving higher; because of that, I'm choosing to be somewhat cautious with the near-term figures. Even so, my outlook on what lies ahead is extraordinarily positive. Two things drive that optimism, and both connect to artificial intelligence. The first one is Autopilot, the AI agent and orchestration layer we position directly alongside our customers' work as they interact with consumers. The second one is the agents we're creating within Blend, which are beginning to handle work of our own. I believe that together, these 2 pillars open a path for us to achieve 10% to 15% incremental growth on the top line as soon as 2027, along with greater efficiency and speed internally as a company. Let's begin with Autopilot. If you're new to the story, Autopilot is our flagship AI agent. We introduced it and launched it in beta almost precisely 2 months ago, letting our customers use it at no cost and test it throughout all of Q2 so they could watch it operate and help their business. As of Monday, May 4th, 65 lenders have turned Autopilot on, 22 are running it live in production and more than 7,000 applications have already been touched by Autopilot since we went to live production. Their early results, we're seeing, are getting better, both on cycle time and on conversion rate. Two of our biggest lenders are in the middle of implementing Autopilot right now, with go-lives scheduled for Q2, and we have 3 additional top-20 logos in our net-new pipeline for which we expect Autopilot to serve as a significant catalyst in getting those new logos closed. Altogether, we already have $10 million in pipeline sitting there because it addresses a genuine problem for our customers and the consumers they serve. Yet the bigger story, as far as I'm concerned, and for our company and our customers and our shareholders, is just how fast that product keeps developing. We've been posting details to our blog each week, and there are 2 I want to highlight. The first one is Autopilot chat. We rolled that out roughly a month ago, a conversational interface letting a borrower ask Autopilot questions about their loan in everyday language while they move through the process. Which documents are still outstanding? Why are you requesting this particular item from me? Why does it matter given my circumstances? What comes after this? Rather than handing someone a fixed checklist or telling them to pick up the phone, a borrower gets genuine situational awareness of what is happening, which guides them through everything. That's the sort of experience people now anticipate, and we're right at the forefront of delivering it. Then there's the piece I find even more thrilling — Autopilot MCP. What it does is unlock the Blend Platform, letting our clients create their own agents sitting on top of Blend, or run Blend headlessly inside the workflows they already use, while still gaining everything: the compliance, the data model, the workflows, every native integration we've built, plus Autopilot's intelligence layer. A big mortgage company among our customers has already put together a voice agent with it, and I see this as hugely significant and very encouraging for clients who want to own an expanding share of what they're capable of doing — while moving at real speed. That dynamic — customers innovating alongside us and around us, not in place of us — is precisely what we're after and precisely what I anticipate seeing a lot more of down the road. Taken together, all of this amounts to something I consider genuinely powerful. Our customers can now trace a route from the first borrower contact clear through to clear to close with no team member ever needing to touch the file; they're still able to work the file, of course, but it's no longer mandatory. That's a fundamentally different kind of value from anything we could provide previously, or that the industry could provide, and it's something I fantasized about offering back when I founded the company in 2012. Agentic AI has now turned that fantasy into reality. Beyond that, just 8 weeks in, we're releasing at a pace that Blend years back, and nearly every enterprise software firm, would think of in quarters. Each of those updates is rooted in what our customers require, what they say they're after, and how we can affect and better their business; with adoption already rolling, let me update you on our monetization approach. Up to now Autopilot has been in preview, with our focus on bringing real customers live and demonstrating value. Come the end of June, we'll shift into paid tiers. Naturally, as with any contemporary software company, certain baseline capabilities will simply be part of our workflow and deliver intelligence — for example, did you upload the correct document. That's helpful, sure. It'll reduce some friction for consumers starting out and grasping AI; yet the paid tiers are where the complete product resides — what we call underwriting intelligence — where Autopilot reads documents, takes genuine action on the loan file, performs calculations, reconciles against guidelines and pushes the work ahead. Eventually, what we're aiming for is to shift Autopilot's paid tiers into a per funded loan structure, the same way the remainder of our mortgage suite works. That's the correct structure for the long haul, and our customers appreciate it since it lets them view and monitor value loan by loan, while we only get compensated once they close a loan successfully. So for us, that's an excellent product. It creates strong alignment between us and our customers, and it pushes us to guarantee that this delivers genuine improvements in funded value at the loan level. If Autopilot enables a lender to fund additional loans while keeping staff numbers flat, then our revenue grows alongside their achievements rather than their employee count. That's the approach we've used to build Blend from the start, and in an agent-first world it matters even more now. As additional customers come aboard, we'll keep sharing Autopilot updates, yet I want investors to grasp that for us this isn't some minor incremental line item. On top of the already-growing mortgage and consumer banking suites that we intend to keep expanding, Autopilot represents an entirely new leg of growth for the company.

Nima Ghamsari executive
#3

Prior to turning to Autopilot, I'd like to take a moment on a topic I consider genuinely significant and that investors keep raising with me. Where does durable value in enterprise AI truly accumulate? That's the billion-dollar question. This remains an active discussion, and it matters to grasp where Blend stands and how I view it. Over the past couple of years, both the industry and the broader world have concentrated on foundation models — which one is fastest, smartest, tops the benchmarks, and costs least — and that concentration makes sense; however, as capabilities across models converge and innovation continues, durable value is migrating up the stack toward the orchestration layer sitting between the model and the workflow, into what's commonly termed the harness, the element generating real end business outcomes. To state it plainly, the harness is a system that directs the engine plus every surrounding tool toward a dependable, governed result, which is enormously important in a sector such as ours, financial services; meanwhile, the data, the documents, and the particular context of any given moment serve as the fuel that renders all of this genuinely useful. Autopilot is precisely that; it isn't a model. Autopilot draws on the strongest models available beneath it. Rather, it's the orchestration layer that determines what action to take given that precise point in the loan, pulling the specific guidelines, obtaining complete loan context, executing the correct calculations, checking outputs against investor and regulatory requirements, refreshing the loan file, and setting off the native Blend workflows that advance the file. Such logic is tied to that particular loan and that particular consumer in front of it, and it's the sort of work generic AI wasn't designed to handle. A system is required around it, and that's precisely where Autopilot comes in. Autopilot MCP simply elevates this further; it lets users of the Blend Platform construct their own agents, or even work with Blend in a fully headless manner, meaning the harness turns into a platform enabling them to move very quickly since they receive all the regulation, compliance, integrations, and Autopilot intelligence ready-made, and they can create their own experiences and agents on top of it, which represents a substantially different degree of importance because you then become more of the engine — the powered by — rather than the interface. That's where agents can prove truly powerful. This compounds further as we unlock additional capabilities so customers can build more rapidly atop us; that's why my confidence about Blend's position in the AI landscape grows each quarter. For origination, we are the vertical industry harness. We possess the proprietary data that makes that harness function. Our existing business model positions us well to seize what automation offers, while still delivering most of that value to the customer and, we hope, the consumer — and that's the sustainable spot to occupy. That's exactly why I'm enthusiastic about where Autopilot stands. So yes, we hold a bullish view on pillar one, agents for our customers; yet my confidence runs even higher on another front — our own internal efforts, the way we deploy agents there. Over the past several months, we've been developing something we blandly refer to as Blend Background Agents. The concept isn't novel, yet it's straightforward. Whenever an input arrives from outside, whether a ticket, a customer issue, or a feature request, ahead of it landing with a team member, we want an agent to handle the initial pass and act on it. Then the team member looks it over and gives approval. In practice, imagine a ticket arrives describing a bug in our system. An agent instantly grabs it from our support queue, examines it, pinpoints the bug, drafts the code that fixes it, runs tests confirming the bug is resolved, then hands it to a person with the message, "Hey, I needed to change these 10 lines to 50 lines of code, can you approve this?" That shifts our team away from manually steering the car, making the turns, and working out how to travel from A to B, toward playing air traffic control over, hopefully, dozens of cars. To enable this, our agents now have access to our internal tools, our whole code base, and the capacity to spin up environments. They now handle the first pass before our engineers or support team ever lay eyes on that issue. Looking at the figures, the new approach to how we're bringing AI into Blend has already produced over 1.5x productivity in 2026 compared with 2025, measured by the number of pull requests our engineering team completes. And we're only at the beginning of that. Prospects and customers are already noticing how quickly we're moving. Notes from customers reach me constantly. I've visited our biggest customers in person during the last month, and I can tell you the momentum is palpable. Our customers have picked up on a shift in our quality and speed. Let me make one thing clear—what we're talking about isn't some trial run confined to a single team. The very same approach, where agents handle that initial round of tasks, ought to be implemented across all positions at every organization, and in Blend's case specifically, it will be implemented in the roles we have right here. That might look like bringing a new client aboard, getting ready for a customer business review where we travel to their location, or perhaps something as obscure as receiving a manual Excel spreadsheet listing which loans have been funded so our accounting department can process it before anyone on that team has to touch it. During the previous call, I stated our goal of landing among the top 1% of companies worldwide when it comes to adopting Agentic AI. That wasn't an idle remark—we intend to make it happen. This is a topic I care deeply about, and we'll continue pushing toward that objective. Once this is accomplished, my belief is that these efforts, alongside Autopilot, have opened the door to 10% to 15% additional top line growth, plus far greater efficiency and speed for our organization. That speed is likely the single most critical factor for any business, and particularly for an organization such as Blend. It translates to more customer problems resolved, more excellent features built, more initiatives along the lines of what we've accomplished with Autopilot, continued expansion of Autopilot, quicker quarter closes, improved readiness for customer business reviews—these will define the new Blend. In closing, converting a business of our scale into an agent-first operation is unquestionably harder and more intricate than most people realize, yet it's worthwhile. We hold a truly significant mission. The clients we serve reach millions of consumers nationwide annually, so this transformation can't arrive soon enough.

Nima Ghamsari executive
#4

We're moving on this as quickly as possible, and honestly, from where I sit, I think we're the best positioned company in this space. This is something I devote a great deal of my time to, and the team is even more passionate about it than I am. So although the war, tariffs, oil, and all of that may force some caution regarding near-term mortgage market figures, since the macro picture and the rollout timeline for what we're building could also take a while, I've never felt more energized about the medium term, and hopefully the long term too, for our customers, our team, and our investors. With that, I'll hand it over to Jason to walk through the financials.

Jason Ream executive
#5

Thanks, Nima, and I appreciate everyone else being on the call with us; we kicked off 2026 strongly, with revenue and non-GAAP operating income each landing above the top end of the ranges we had guided to. Our top line was up 15% versus a year earlier, and non-GAAP operating margin widened to 13%, which speaks to growth throughout the business as well as the operating leverage we keep embedding into the model. First-quarter 2026 total revenue came in at $30.8 million, exceeding the upper bound of our guidance, with gains coming from both Mortgage and Consumer Banking. Mortgage Suite revenue reached $17.2 million, a rise of 18% from the same period last year. Roughly $187,000 in loans were funded through our platform during Q1, which is 29% higher than a year ago and a touch above what we had anticipated entering the quarter. That robust volume expansion was partly counterbalanced by a decline in economic value per funded loan on a year-over-year basis, landing at $84 in Q1, which sits inside the $84 to $85 band we mentioned during our prior call. We're at the bottom of that band, largely due to elevated mortgage volumes, which by its mechanics pushes down the per-loan economic figure because of certain fixed fee structures present among our customers. In Q1, our Consumer Banking Suite brought in $10.8 million, which is up 12% compared to the same period last year and matches what we told you about on the prior call. Professional Services generated $2.9 million in Q1, up from $2.1 million in Q4 on a sequential basis. Of that $2.9 million in Professional Services revenue, roughly $600,000 came from projects completed in prior periods that we recognized this quarter under our revenue recognition rules. We don't expect a similar catch-up amount going forward. Turning to profitability: non-GAAP gross profit was $24.8 million, with a non-GAAP gross margin of 80.3%, up from 72.9% in Q1 of 2025. I'd point out that gross profit this quarter got a lift from the PS catch-up I referenced a moment ago, plus a one-time cost of revenue benefit, and together those items raised gross margin for the quarter by roughly 2 to 3 points. Kindly bear that in mind when considering how to model gross margin going forward. Non-GAAP operating expenses totaled $20.7 million in Q1, which is 10% higher than a year earlier. As a reminder, this year-over-year comparison incorporates the shift in our internally developed software capitalization methodology that we covered last quarter, under which we capitalize a smaller portion of our R&D personnel costs than was the case in 2025. That's a change in accounting treatment, not a change in how we approach our R&D investment. Consequently, reported R&D appears inflated relative to last year, and that effect will linger somewhat across 2026 until we anniversary the prior-year periods. Non-GAAP operating income was $4.1 million, topping the high end of our $2 million to $3 million guidance range and equating to a non-GAAP operating margin of almost 13%, roughly 10 points better than the first quarter of 2025. During the quarter, free cash flow came in at $7.3 million, versus $15.5 million a year earlier. We feel good about how strongly we're generating cash, and I'd like to point out our seasonal tendencies — in our business, Q1 tends to be a robust quarter for collections. Our balance sheet also continues to be in a solid position. We closed the quarter holding $59 million across cash, cash equivalents and marketable securities, along with 0 debt. To put our cash to work, we bought back 11.2 million shares this quarter, paying an average of $1.66 per share through our share repurchase program, which meant deploying $18.6 million out of the $50 million authorization we told you about on our previous call. Three months back, we told you this initiative reflects both our belief in the company's value over the long haul and our commitment to allocating capital with discipline. With no debt and solid liquidity, our balance sheet lets us put money into the business while also returning it to shareholders. Before I get to the outlook, let me touch quickly on market share and the broader economic picture. Starting with market share: preliminary 2025 HMDA figures, published in early April, pointed to about 4.4 million originations across the year, which puts our 2025 mortgage market share near 17% — squarely within the 16% to 18% band we set as guidance last November. That HMDA figure will keep settling as late filings arrive, though we don't anticipate it shifting meaningfully. Looking ahead to 2026, we anticipate a market share headwind of roughly 100 basis points, mainly reflecting the volume roll-off from one large customer we've previously discussed. Right now, we don't see any other meaningful headwinds affecting our market share. On the macro front, the spring housing market got off to a stronger start than many anticipated, helped by improving affordability and inventory that is slowly being rebuilt. Even so, the recent uptick in mortgage interest rates introduces uncertainty into the outlook. According to Fannie Mae's most recent projection, the mortgage market overall should expand by roughly 19% compared with a year earlier in 2026 — although earlier this month, with rates moving upward, Fannie trimmed its outlooks for both the second quarter and the entire 2026 year. That revised Fannie forecast serves as the basis for our own 2026 perspective. Until we see a meaningful drop in rates and a pickup in Refi activity, we intend to keep our outlook conservative. Even so, our platform and customer base are ready to seize the upside when conditions get better. Let's move on to guidance now. Looking at the second quarter of 2026, we're projecting total revenue in the range of $32 million to $34 million, which works out to roughly 1% to 7% growth versus the prior year. Looking past the top-line figures, our projection for Mortgage Suite revenue is an increase of 4% to 10% compared with the same period last year. Growth in mortgage market volume is the force behind this, though it will be partly counterbalanced by economic value per funded loan falling relative to a year ago — for Q2, we anticipate that figure landing somewhere between $79 and $80. The decline in evPFL from Q1 into Q2 is largely attributable to greater volume, and, as I noted earlier, that naturally drives evPFL lower. For Q2, our expectation is that Consumer Banking Suite revenue growth year over year lands somewhere in the range of negative 2% to positive 4%; meanwhile, Q2 non-GAAP operating income should come in somewhere between $5.5 million and $6.5 million, which works out to a non-GAAP operating margin of roughly 18% at the midpoint. Let me add a handful of points about what sits inside those expectations. Macro volume swings keep affecting our Mortgage Suite business. And how mortgage rates and the wider housing market move from here will determine whether Mortgage Suite revenue moderates or perhaps even goes flat during the back half of 2026, especially if Refi activity stays weak. Now, on the economics of each loan: Q1 tends to be the peak because of seasonal patterns, which explains why our guidance calls for a decline between Q1 and Q2 — going from $84 in Q1 to somewhere between $79 and $80 in Q2. We aren't factoring in any boost from Autopilot, since it's too soon to put a number on that and it isn't part of any of our expectations. Absent that, we'd expect evPFL to move with seasonal patterns during the back half of 2026, though it would stay below where Q1 came in. Turning to Consumer Banking, growth is slowing down as a result of the headwinds we mentioned during our previous earnings call. On top of that, home equity has seen weaker volumes tied to the macro environment as rates have moved higher. Combining those two elements, our forecast for Consumer Banking in the latter portion of 2026 is year-over-year growth in the single digits, and Q3's figure will likely land under Q4's due to how the comparisons stack up against last year. On top of that, macro conditions do affect the home equity segment within Consumer Banking. As a result, if rates were to move higher from current levels, additional strain on those growth rates is what we'd anticipate. Finally, a few words about Autopilot specifically. The possibility that Autopilot drives revenue upside has us very enthusiastic, though at this point we would advise investors against being too aggressive in building it into their models. Once the free trial phase is finished and we've gained a bit more experience, we intend and hope to share further details regarding possible effects on the outlook. In summary, as we move through the remainder of 2026, we're quite optimistic about how the business is positioned. Mortgage recorded its second straight quarter of growth compared with the prior year in Q1; since churn has settled down and the shift to the partnership model is now complete, we anticipate that from this point forward, macro factors will be responsible for most of the fluctuation in Mortgage revenue. We've kept a firm grip on expenses, and throughout the year we anticipate generating further gains in productivity and efficiency, since AI-powered workflows keep building on each other within our internal operations — work that, as Nima mentioned, is already moving forward company-wide. Truly, this marks a thrilling moment for Blend, and we trust you feel that excitement about joining in as well. With that said, we'd now like to turn the call over to your questions.

Ryan Tomasello analyst
#6

Ryan, thanks for the question. In the remarks I prepared, you heard me say that Autopilot and our AI efforts open up a route toward, I believe I said, 10% to 15% additional top line growth. Could you sharpen what that actually means, and what gives you confidence to put a number on those benefits this early?

Nima Ghamsari executive
#7

Sure. Good to hear your voice, Ryan. What I'd point to first is the pipeline we have today. Right now our Auto pipeline sits at roughly $10 million; we've been out in the market with pricing for our customers for a little more than a month, and many customers have switched it on, with really encouraging feedback coming back to us. I brought up 2 — 2 very large go-lives with customers. So provided we sustain that momentum, my framing is 10% to 15% incremental on top of whatever other growth you might be modeling that comes from Autopilot — that's the path we see at present. Naturally we still have to keep executing; plenty of work lies ahead of us, but the product is awesome and our customers love it.

Ryan Tomasello analyst
#8

Great. Maybe we shift over to Consumer Banking now. With all the disruption caused by that significant customer loss, could you help us get a sense of what the underlying revenue growth rate looks like for that business? For 1Q as well — and then more broadly, going off the figures you shared earlier, I believe there was a $2.5 million hit tied to that big client in Consumer Banking. The growth trajectory there just appears to be landing somewhat softer than what people had originally expected. So Nima, any wider thoughts on how you're feeling about that segment's momentum moving forward?

Nima Ghamsari executive
#9

Yes. I'd agree — the largest effect comes from that big customer, and they carried a fairly substantial Consumer Banking line item, which you pointed to. That said, on the brighter side, we've got several quite sizable financial institutions that will be launching our wall-to-wall suite during this year. Those implementations are underway. So that has us enthusiastic. When that lands, I believe it will deliver a favorable lift. Beyond that, we also have excellent clients bringing our rapid home equity product to market right now. That represents yet another — it will serve as yet another favorable catalyst once it occurs. Naturally, the home equity sector has additional macro factors at play too. Still, there are sufficient fresh developments occurring across the Consumer Banking space generally that leave me feeling quite confident about that business.

Dylan Becker analyst
#10

Nima, I appreciate all the detail you gave on Autopilot and Autopilot MCP, I suppose. Clearly, it sounds like plenty of customers out there want to try piloting this. I believe you mentioned a few early proof points involving better cycle times and conversion rates. Could you maybe shed a bit more light on how that compares with a process that isn't automated, just to concretely quantify what customers are experiencing and discovering? And then perhaps also how you're viewing deployment or usage of first-party agents against the MCP-enabled agents, and maybe the economic differences between them.

Nima Ghamsari executive
#11

Yeah. Regarding the impact, I'll share 2 anecdotes involving 2 customers who've been among the heaviest users of it; we assist them in tracking cycle time and conversion. Why conversion improves is less obvious. So I've actually spoken with one of our customers about this. I'll cover that shortly. Take cycle time at one of those customers, for instance, going from application completed in Blend to closing docs delivered to the customer — or closing disclosures, rather, delivered to the customer — it dropped from 29 to 21 days at one customer we had. That's a fairly significant gain in their cycle times. And I think it fundamentally makes sense, since customers go back and forth with consumers a great deal. What Autopilot does is, in real time, while the consumer is in the flow, it spots those items that will become problems further down the line; it tells the consumer, hey, we see this account is held in the name of a trust, we need your trust documentation immediately rather than requesting it several days later after an underwriter reviews it and passes it to a processor, who then sends it back to the loan officer. So it effectively shortcuts the process in a good way to let the — our hope with Autopilot plus some of the Rapid products, when you combine those 2, I'll call it RapidPilot, is that you can get an application started and approved, because Rapid provides an approval and an offer upfront. Then, once that customer is set to proceed, get them clear to close within minutes; that's the world we aim to enable for our customers and consumers, or possibly conditionally clear to close pending an appraisal if an appraisal is required for a mortgage. So I think that's a — those are some encouraging figures. I think what's surprised me more is why conversion is so much stronger. Though I suppose it stands to reason that if you can provide folks with greater certainty more quickly, we're also observing solid improvements in conversion, which clearly it's still early days, yet that matters even more to our clients since those are borrowers who would otherwise leave and whom you, as a lender, would invest both time and money in—not merely for items such as credit pools and other data retrievals, but additionally the effort and energy your staff put into it. Thus, if we're able to reduce these timelines and turn the lending workflow into something closer to instantaneous, I believe it genuinely reshapes the sector. And one point I'd like to make regarding Consumer Banking, since right now we're working on developing the connections to every Consumer Banking offering for Autopilot. I believe there's a chance there. At present, Consumer Banking contains fewer tasks done by hand, yet the sheer quantity of such tasks is far greater when it comes to how many units these clients complete; and therefore although it might not justify thousands of dollars per loan for our clients over the long haul in Consumer Banking or for each new account, the magnitude of these operations genuinely matters to them, and they maintain very large operations departments overseeing these workflows; so it's a significant capability for them to be able to process far more volume using those departments, and I believe Autopilot makes that possible. Another item. Yes, this just came to mind. The additional factor that I believe has long been a challenge, and I brought this up during my prepared remarks, is that rates are—they truly propel refinancing activity above all. Consequently, if you're a mortgage servicer whose volume is heavily concentrated in refi, the sole method available to you for coping with massive volume is expanding and contracting—in the past it has meant growing and shrinking your teams. And we simply cannot reliably forecast when rates will decline. I mean, you attempt to, but forecasting when the upcoming fresh figures will be released, or what those figures will show, is extremely difficult. And therefore the capacity to build flexibility in staffing when you now possess agents that a lender can activate and deactivate in conjunction with their own staff before that staff performs the work—with the agents handling an initial pass—I mean it simply alters the economic characteristics of servicing and the recapture of servicing. I mean, I think for our major servicing clients, of whom we have plenty, it's going to shift—I believe it's going to shift how they can conduct business, since it will let them manage these market swings even more effectively than—even more crucially for them than somebody operating on the purchase side.

Dylan Becker analyst
#12

That makes sense. And then perhaps for Jason, following up on that as well, you sort of highlighted the dynamics around per funded loan and market share. I would love to dig into that once more, if you could remind us — it sounds like you really are growing market share thanks to customer momentum and the customers who are coming online or getting onboarded, but that is sort of pulling in the opposite direction near term when it comes to per funded loan volume. So I suppose walk us through the mechanics behind that, and also perhaps when we should expect that to reverse, with those 2 tailwinds maybe operating together or side by side, so that you see both an inflection in market share growth and an expansion in per funded loan over time.

Jason Ream executive
#13

Yes. Good question, Dylan. I believe we are observing volume growth — as I said, our volume in Q1 was better than we had anticipated even when we entered the quarter. Partly that reflects our customers performing better. Partly the market was somewhat stronger than we had anticipated during the quarter. And then naturally, yes, we are constantly attempting to gain share and bring new customers onto the platform. When it comes to the per funded loan, setting aside the seasonal variability that arises from the mechanics I described, we are, I believe, making a far more focused push right now to generate growth year-over-year with customers we already have. I think that — look, items such as Autopilot provide us with improved pricing leverage when entering new customer situations. Clearly, that generates its own revenue stream, but it additionally gives us leverage in the core platform, so to speak. And then I think Rapid continues to be a driver as well on the refi side, especially; and as Nima said, refi is even more responsive to rates than purchase. And so — and we do not offer a rapid purchase product. What we have is a Rapid Refi product. And so as rates decline, we ought to see a benefit in volume and revenue in that regard, but also as more customers come onto Rapid Refi, see a benefit in PFL as well.

Joseph Vafi analyst
#14

Nima, could you give us an update on how Rapid is being adopted and what kind of reception you're seeing in the market? Clearly, conditions out there aren't as favorable as we'd hope, but tell us some of what you're hearing; then I've got a follow-up for you.

Nima Ghamsari executive
#15

Sure. I'd restate my earlier point regarding RapidPilot — that's the piece where customers seem to be directing their energy and attention, meaning Rapid combined with Autopilot. That's where a great deal of my time goes. Over the past 2 weeks, I've been on-site twice with 2 very large banks and lenders discussing precisely this, since they're aiming to have it live in Q2 — because in reality, our customers, particularly when it comes to refis and home equity, want the ability to present an offer immediately and then complete all the necessary work behind that offer immediately as well; the fusion of those 2 capabilities has proven remarkably effective. Beyond that, as I mentioned when answering a question earlier, several very, very large customers are going live with Rapid home equity, among them some of the leading home equity originators nationwide; so it's certainly a favorable moment for the industry. If I were to level one criticism at myself, it would be: how do I make adoption so simple that they flip a switch, turn it on, and Rapid Refi is now enabled in their environment? That, I believe, is a challenge on our end — something we're contemplating over the coming couple of months; and we're determined to bring it about. Yet it's also something that, once we do, will let our customers adopt it far more easily. In fact, that proved a key lesson from our Autopilot effort and rollout, where we made turning it on genuinely self-serve for a customer, and we're witnessing the adoption. Consider the figures we provided regarding how many lenders have switched this on relative to our total, but just imagine large financial institutions activating a new AI agent across their organization with a flip of a switch, without so much as calling us. What surprised me most was that fairly large banks — very large banks — were enabling this in beta and production without our even being aware. Then we discovered it, saw it begin streaming through our logs, and thought, "Oh, we ought to probably get in touch with them and discuss it." That's not to claim we're a product-led growth company. We do enjoy speaking with our customers so they get the most out of what we offer. However, if adoption is simple, that will benefit Blend tremendously, and it all ties back to velocity—both how quickly customers take it up and how fast our team can iterate. Assuming we can pull that off—and I have a lot of faith that we demonstrated this capability with Autopilot—then I believe we can carry that smaller-scale mindset and smaller-scale offering, along with those ideas, into every other area of Blend’s business, and that would let us help reshape the whole terrain of how these offerings get used and what kind of difference they make for our clients. Let me close with a final story, which isn’t precisely what you asked about, but Autopilot MCP has created many new possibilities for us. Last week I visited a rather sizable customer, and their head of engineering was present. The very first thing that head of engineering wanted to know was whether they could incorporate this into their mobile app, and I told him absolutely, there’s now a method for that: Autopilot MCP. Every Blend capability plus Autopilot’s intelligence layer can live completely inside your own setup. He reacted with surprise and said okay. Then he posed what I thought was a really strong question: could he apply it elsewhere in his business? For those other loan types—and you mentioned a few other loan types—they don’t use Blend. I said yes, certainly, Autopilot functions there. You’re able to add bespoke guidelines on your own, without any need to contact us, and he got visibly excited. While we were still in that room—a rather large room—the first thing he requested was a copy of the Autopilot document, meaning the Autopilot MCP documentation, and we sent it over to him right then. So these are folks who previously had a hard time figuring out where they fit within the Blend ecosystem, or how to slot their tech stack into it, and today we’ve made that possible. We also had a really fascinating sales conversation with a rather sizable bank, and their digital leader joined, someone who historically can feel somewhat sidelined by us when we get brought in. His opening question was whether he could use this alongside his existing digital stack. The moment the response came back as yes, obviously, now that Autopilot MCP exists, he shifted from likely being a critic to someone saying, oh wow, this is genuinely quite compelling. Now I’m able to deliver new digital capabilities. I can help enhance my customer experience in a powered-by fashion that would otherwise require months, possibly years, for them to accomplish in-house, plus building agents that are this capable and this intricate. Joe, I realize that wasn't quite what you asked, but while I was speaking, that memory just came back to me.

Joseph Vafi analyst
#16

Yeah. That's definitely a thrilling setup. I appreciate the context, and I'm eager to see how things move forward there.

Aaron Kimson analyst
#17

Nima, based on what you've been hearing, what value do customers see Autopilot delivering right now? Is it still mostly regarded as something that comes out of tech budgets? Or are financial institutions starting to be more willing to treat agentic offerings such as Autopilot as part of what they spend on labor?

Nima Ghamsari executive
#18

That's a fascinating point. At this moment, I'd say businesses are still working this out, even as we talk. They can't yet answer that precise question you're asking. That's really why, for the near term, we've set our pricing so that our customers can try it out for a couple of months at no cost. Yet beyond that point, we'll have essentially level pricing, which is clearly beneficial for us from an economic standpoint, while also helping our customers by giving them near-term breathing room to make the necessary adjustments to their processes and their teams; so — that said, over the longer horizon, everyone agrees on one thing: labor no longer has to be expanded or reduced in step with volume. I recently spoke with the chief executive of one of our biggest customers. The prospect of growing their company without bringing on thousands or more employees is incredibly powerful. So, in the end, it becomes a labor-related matter quite naturally. Still, I'd argue the bigger value lies in the fact that as these figures on conversion rates become firmly established and we understand them better, that will deliver even greater benefit to our customers, since a huge number of consumers across this country stand to gain from reduced interest rates, home equity, debt consolidation, or any of these other things that are going on, which historically have been difficult for our customers to take advantage of, and difficult for consumers too, because the process is so drawn out. So if we can now make everything fully clear through something like Rapid, and then fully automate it through something like Autopilot, that will transform things — it will mean far less friction for consumers in this process, which means more of them will pursue it, and they'll do so through our customers.

Aaron Kimson analyst
#19

Understood, that helps. One last thing. You've partnered with banks for quite a while. Could you discuss why they seem more willing to take on new offerings quickly now compared with before? Also, how do they weigh building in-house against purchasing, and how do they split between bringing in AI tools and backing AI-first startups rather than going with proven software companies such as Blend — plus, where do the Frontier Labs come into that? I believe all of us are attempting to sort this out when it comes to application software broadly.

Nima Ghamsari executive
#20

Right. I'd say we're in a rather curious spot where something shifted at some point during Q1 of this year — around February, I'd say — when our clients began to grasp, possibly driven by the whole Anthropic Claude Code surge that's taken place across the industry, just how significant a change this would turn out to be, and every one of them has allocated budget toward AI and AI efforts since they recognize the stakes. It matters for the people they serve. It matters for the users on their platforms, and it truly matters for the business's economics over the long haul. This stuff can accomplish some very potent things. And I believe folks are beginning to accept that; it stopped being something they regarded as belonging to a future 2027, 2028 scenario. Instead it became, well, I can genuinely pull this off right now. So I'd say that largely makes the case on its own, along with the plain fact of how many of our big financial institution clients have switched this on — activated these features themselves — and are either actively talking with us or already in the middle of a process to deploy these features at scale. Still, yes, I'd say that shift occurred at some point earlier in this year. And they do tend to reason through how this slots into our — into the stack they use. Is it an organization such as Blend, which already powers a great deal of their internal work and the work they do for clients? Or instead — are they engaging Anthropic or OpenAI or another firm through something resembling a large consulting-style engagement? Or are they going with a tiny startup? My take on Blend against — or rather Autopilot against a tiny startup, is that since such a large portion of the workflow already runs inside our platform, which gives natural points of entry for calling up AI agents and then shutting them back down, we hold a fairly solid edge in helping our clients move really fast along those lines; so our task is ensuring Autopilot stands as the top offering available for precisely the kinds of tasks our clients must handle; and here, perhaps that means underwriting intelligence, as I mentioned during the prepared remarks. And provided we keep doing that, my sense is they won't switch to a tiny startup. We simply have to keep pace, and we are keeping pace. Also, a strong product has to be created, and Autopilot is exactly that kind of strong product. It is accomplishing things that, 12 months ago, our customers would have regarded as science fiction. So, my view is — and separately, there is a conversation about the way they consider the labs compared with the big labs such as Anthropic and OpenAI, as opposed to a company like Blend. And I believe part of that is still uncertain. I have heard about truly impressive work the labs are doing with many of our customers. And I believe a huge portion of the industry will transform. The pie is likely far larger than anyone truly grasps. And the labs will not step in and attempt to create something within our workflow so they can generate value for our customers. I mean, I do not believe they would. But even if they did, we are already established there. We already possess it. And so speed matters greatly when it comes to adoption. And if you must spend a 9-month or 12-month project to obtain something rather than being able to flip a switch, making that possible is our job.

Operator operator
#21

That brings us to the close of the question-and-answer portion. Today's call ends here. Thanks to everyone for joining. You're free to disconnect at this point.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Blend Labs, Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Blend Labs, Inc. earnings transcripts and 131,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.