Blend Labs, Inc.

Hello, everyone. Thank you for joining us and welcome to BlendLab's first quarter 2026 earnings call. After today's prepared remarks, we will hold a question and answer session. If you would like to ask a question, please press star one to raise your hand. To withdraw your question, press star one again. I would now like to hand the conference over to management for prepared remarks. Please go ahead.

Good afternoon, and welcome to Blend's financial results conference call for the first quarter of 2026. I'm Meg Nunnally, Blend's head of investor relations. Joining me today is Nima Gamsari, our co-founder and head of Blend, and Jason Rehm, our head of finance and administration. Before we start today's call, I'd like to note that we 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 second 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 laws. 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-K for fiscal year 2025 and our 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 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. It's been a whirlwind two months since our last call. We reported our Q1 numbers today, which Jason will spend time on, but came in higher on revenue and non-GAAP operating income than expected. We also signed 15 new deals and expansions in the quarter, including an e-closed deal with a top 20 bank, along with a new mortgage deal with another top 100 bank. Our pipeline as of March 31st is up more than 40% year over year, and that doesn't include the autopilot pipeline, which I will cover in a minute. But the world has shifted underneath us in those two months. Increased global conflict, inflation, and a rise in mortgage rates. And that leads me to be a little conservative in the short-term numbers. but I am incredibly optimistic about the future. My optimism comes from two things, and they are both tied to artificial intelligence. The first is autopilot, which is our AI agent and orchestration layer that we put right alongside our customers' work as they work with consumers. The second is the agents we're building inside Blend, which are starting to do our own work. Together, I believe these two pillars give us a path to see 10% to 15% incremental growth already for us in 2027 on the top line and more efficiency and speed as a company internally. Let's start with Autopilot. For those new to the story, Autopilot is our flagship AI agent. We unveiled it and rolled it out in beta almost exactly two months ago, telling our customers they could use it for free and try it out for all of Q2 to see it in action and help their business. As of Monday, May 4th, 65 lenders have activated Autopilot, 22 are running it live in production, and over 7,000 applications have already been touched by Autopilot since we moved to live production. And we're seeing their early results are improving, both in cycle time and in conversion rate. Two of our largest lenders are actively implementing Autopilot right now with GoLive's plan for Q2, and we have three more top 20 logos in our net new pipeline that we expect Autopilot to be a meaningful catalyst for closing those new logos. In total, we're already sitting on $10 million in pipeline because it solves a real problem for our customers and the consumers they serve. But the more important story for me and for our company and for our customers and our shareholders is how quickly that product is evolving. We've been publishing details to our blog every week and there are two that I want to call out. The first is Autopilot Chat. That was rolled out about a month ago, a conversational interface where the borrower can ask Autopilot questions about their loan in plain language as they're going through the process. What documents are still needed? Why did you ask me for this specific thing? Why does it matter to my situation? What happens next? Instead of a static task list or making a phone call, the borrower can have a real contextual understanding of what's going on to help them through the process. This is the kind of interaction that consumers are starting to expect, and we are right on top of it. The second is something I'm even more excited about, which is Autopilot MCP. That opens up the Blend platform so that our customers can build their own agents on top of Blend or use Blend in a headless way in their existing workflows and still get the benefit of all the compliance, all the data model, the workflows, all the native integrations we built, and the intelligence layer of Autopilot. One of our large mortgage company customers has already built a voice agent using it, and I'm seeing this really important and really promising for our customers who want to own more and more things they can do, but to move really fast. And that pattern, customers innovating with us and around us rather than instead of us, is exactly what we want and exactly what we expect to see more of going forward. What this all adds up to is something I think is really powerful. Our customers can now see a path from initial bar of a touch all the way to clear to close without a team member ever having to touch a file. Now, they still can work on the file, but they won't have to. That is fundamentally different value than we could ever offer before or the industry could ever offer. And something that I dreamed of being able to offer when I started the company in 2012. And now Agentic AI has made that dream possible. And on top of that, eight weeks in, we're shipping at a cadence that blend of years ago and most enterprise software companies would measure in quarters. And every one of those updates is ground on what our customers need what they're telling us they want, and how we can help impact and improve their business. With adoption well underway, let me give you an update on how we're going to monetize this. Autopilot has been in preview to date, and our priority has been getting real customers live and proving the value. Starting at the end of June, we're going to move to paid tiers. Now, just like any modern software company, there's going to be some base capabilities just built into our workflow that are going to provide intelligence, like did you upload the right document? But And that's useful, that's going to lower some friction for consumers to get started and understand AI. But the paid tiers are where the full product lives, what we call underwriting intelligence, where Autopilot is reading the documents, taking real action on the loan file, running calculations, reconciling its guidelines, and driving the work forward. Over time, our intent is to move the paid tiers of Autopilot to a per-funded loan model, just like the rest of our mortgage suite. So write long-term structure, And our customers like that because it allows them to see and track the value on a per loan basis. And we get paid when they make a successful loan. And so that's a great product for us. It's a great alignment for us with our customers. And it incentivizes us to make sure this is providing real loan level funded value improvements. When Autopilot helps a lender fund more loans with the same number of people, Our revenue scales with their success, not with their headcount. And that is how we've always built Blend, and that's even more important today in an agent-first world. We're going to continue to provide updates on Autopilot as more customers sign on, but I want investors to understand this is not a small incremental line item for us. Autopilot is a whole new leg of growth for the company. On top of the great mortgage and consumer banking suites that are already growing, and we plan to keep growing it. Before we move off autopilot, I want to spend a minute on something that I think is really important and I keep getting asked about from investors. The billion-dollar question is, where does the durable value in enterprise AI actually accrue? This is an ongoing debate, and it's important to understand where Blend fits and how I see this. For the last couple of years, The focus of the industry and the world broadly has been on the foundation models. Which model is the fastest, the smartest, does the best in benchmarks, the cheapest, and that focus is understandable. But as models converge in capability and keep innovating, the durable value is shifting up the stack to the orchestration layer between the model and the workflow, to the area that people call the harness and the thing that's driving actual end business outcomes. The harness to put it clearly, is it's the system that channels the engine and all the tools around it into a reliable, controlled outcome, which is so important for an industry like ours, like financial services. And the data and the documents and the specific context of any moment is the fuel that makes any of that work actually useful. And Autopilot is exactly that. It is not a model. In Autopilot, we use the best available models underneath And instead, it's the orchestration layer that decides what to do given that exact moment in a loan. It retrieves the specific guidelines, gets the full context of the loan, runs the right calculations, validates the outputs against investor and regulatory requirements, updates the loan file, and triggers the native blend workflows that move the file forward. That logic is specific to that exact loan, exact consumer in front of it. And it's the kind of work that generic AI isn't built to do. It needs a system around it. And that's where Autopilot fits in. And autopilot MCP just takes that to the next level. It allows the blend platform users to build their own agents or even build work with blend in a completely headless way, which means the harness becomes a platform for them to move really fast because they get all the regulation, the compliance, the integrations and the autopilot intelligence out of the box. And they can build their own experiences and their own agents around that, which is just a meaningfully a different level of importance because now you become more of the engine, the powered by, instead of the interface. And that's where agents can be really powerful. And that compounds more as we open up more capabilities for our customers to build faster and on top of us. And that is why I get more confident every quarter about where Blend sits in the AI landscape. We are the vertical industry harness for origination. We have the proprietary data to make that harness work. We have the business model already to help capture the benefit of automation and still give most of the benefit to the customer and hopefully the consumer, but that's the durable place to be. And that's why I'm excited that's where Autopilot is. And yeah, so we're bullish on our first pillar, which is agents for our customers. But I'm even more bullish on something which is what our internal work, how we're using agents there. And so over the last few months, we've been building something that we're boringly calling blend background agents. It's not a new idea, but it's a simple idea. Anytime we get an input from the outside world, it could be a ticket, a customer issue, a feature request. Before that reaches a team member, we want an agent to take the first pass of that work. and do action, take action on that. And the team member reviews and approves it. And in practice, that could be something like a ticket comes in that outlines a bug in our system. An agent immediately picks it up from our support queue, looks at it, identifies the bug, writes the code to fix the bug, tests the code to make sure the bug is now fixed, and then sends it to a human and says, hey, I had to change these 10 lines, 50 lines of code. Can you approve this? And that moves our team from manually driving the car and making the turns and figuring out how to get from A to B to playing air traffic control with hopefully dozens of cars. And so to support that, we've given our agents access to our internal tools, our entire code base, the ability to stand up environments, and they will now take up the first pass before our engineers, our support team ever see that issue. When I look at the numbers, the new process of how we're adopting AI at Blend has already resulted in more than 1.5x productivity in 2026 versus 2025 based on number of pull requests our engineering team is doing. And we're just getting started with that. Prospects and customers are already taking notice of how fast we're moving. I get notes from customers all the time. And I've been on site with our biggest customers in the last month, and I can tell you that momentum is palpable. our customers have noticed a change in our quality and speed. And I want to be clear, this is not a one team experiment. This exact same pattern of agents doing the first pass of work should apply to every role in every company, and specifically in Blend, it will apply to our roles here. And that could be something like onboarding a new customer, preparing for a customer business review, or we're going on site with them, or even something as esoteric as getting a manual Excel worksheet that comes in that outlines what loans have been funded for our accounting team, doing that work before our accounting team even has to pick it up. And I said on the last call that we aim to be in the top 1% of all companies. in terms of agentic AI adoption. And I really meant it. We're going to do it. It's something I'm very passionate about and we're gonna keep driving for that. When done, I believe this effort combined with Autopilot, that's created the path to 10 to 15% more top line growth and a lot more efficiency and speed for us. And that speed is probably the most important thing for any business and especially for a company like Blend. means more customer issues fixed, more great features developed, more things like we've done with Autopilot, continue to grow Autopilot, faster time closing a quarter, better preparedness for customer business reviews. These will be the new blend. To wrap up, transforming a company of our size into an agent-first company is definitely more work and more complicated than the world understands, but it's worth it. We have a really important mission Our customers serve millions of consumers across the country every single year. So this change can not come fast enough. We are taking it as fast as we can. And we feel like to be quite candid from my perspective, the best position company in the space. It is something that I spend a lot of my time on and the team is even more passionate about. So, you know, while the war and tariffs and oil and all those things might have create some conservatism around short-term mortgage market numbers because the macro and the rollout time for building might also take some time. I have never been more energized about kind of the medium term and hopefully even the long term for our customers, our team, and our investors. And with that, I'll turn it over to Jason to walk through the financials.