Blend

🚀 𝐖𝐞𝐞𝐤𝐥𝐲 𝐆𝐑𝐀𝐓𝐈𝐓𝐔𝐃𝐄! 🚀 These three people always ask the right questions, follow through, and ensure our customers are taken care of. From projects, implementations, and to weekly check ins- these are the people who work tirelessly for our customers! 🌟 Elyse Evans you advocate for your customers each and every day. The way you communicate internally and externally is something everyone can learn from!👏 🌟 Yash Gupta our Professional Services colleague working in the trenches on each implementation and expansion initiative. We appreciate everything you do here Blend ! 👏 🌟 Alok Juneja fearlessly leading our Innovation Services team and making sure we are hitting our project dates! Your partnership is invaluable👏 #Gratitude #EmployeeSpotlight #ThankYou

Mid-process product switches, conventional to VA, conventional to FHA, happen constantly in mortgage. The borrower qualifies differently than expected. Rates move. Guidelines change. The problem isn't the switch. It's everything that comes after it: open conditions that no longer apply, tasks that need to be rebuilt under new guidelines, a borrower sitting in limbo while the loan officer manually reconciles the difference. Blend Autopilot now handles this automatically. Change the product, and the agent re-reviews the loan, cancels the irrelevant conditions, and updates the borrower's task list in seconds.

Modernization doesn't have to be a multi-year bet. Start with one product, prove the impact, then expand. That's the approach I shared during our recent webinar with Cornerstone Advisors and People First Federal Credit Union and I think it's one of the more practical ways to get institutions unstuck. Watch the full conversation here: https://hubs.ly/Q04cZ3yX0 Blend #CreditUnions #MemberExperience

The best decision I made on our AI agent rollout had nothing to do with the model. It was putting a deal desk analyst in the scoping sessions from week one. When we started building a contract flagging agent at Blend , the instinct was to keep it in the technical track (engineering, product, me). Standard build process. But I pulled one of our legal folks into the room instead of waiting until UAT. Third session in, wecaught something nobody else would have. The agent's flagging logic didn't account for co-term deals where the renewal date shifts based on the original close date. That's not documented anywhere. It's tribal knowledge from two years of running non-standard deals through DealHub.ai . If that goes live and generates bad outputs on roughly 15% of the renewal book. Not a catastrophic failure. The kind that erodes trust quietly over 90 days until the team stops using the tool and nobody can explain why adoption flatlined. We've since made it a rule: no agent moves from pilot to limited without a subject matter expert from the team that will actually use it sitting in scoping. Not review. Not sign-off. Scoping. There's a version of AI deployment where ops teams get consulted at the end. There's another version where they're in the room when the requirements are written. The second version ships slower. It's also the only one that's still running six months later. #AIDeployment #RevOps #AgenticAI #DealDeskOps #BringInPeopleEarly

Every lender AI demo I've seen this year lives in the application flow. The cycle time dies 20 days later, in condition clearing. Those two things don't overlap and nobody's building for the second one. "Approved with conditions" isn't a status. It's a to-do list that nobody tracks as structured work. At Blend we see this from both sides. With building the LOS and watching lenders operate it. The tech budget lands on application UX, pricing engines, and the point-of-sale. The manual time sits in the middle. Chasing conditions, re-reviewing a paystub rejected for format, re-running AUS every time something shifts. Each condition is its own mini-workflow. Request, gather, upload, review, accept or reject. If rejected, repeat. Most lenders run that cycle through email threads and LOS tasks that never structure the data. That's where AI has real leverage in origination. Document classification, data extraction, condition matching against updated guidelines are all workable. But only if the conditions themselves are tracked as structured entities, not free-text notes in a task queue. The lenders who pull ahead in 2026 won't be the ones with the slickest application flow. They'll be the ones who turned condition clearing into structured workflow first. #MortgageTech #LoanOrigination #AIinFintech

🚀 𝐖𝐞𝐞𝐤𝐥𝐲 𝐆𝐑𝐀𝐓𝐈𝐓𝐔𝐃𝐄! 🚀 The best weeks aren't the ones where everything goes right. They're the ones where the right people show up when it matters. This week I'm recognizing three people who kept deals moving, processes clean, and customers taken care of. 🌟 Lisa Britt way to start off and dive head first in! You have submerged yourself into Blend and so happy to have you here!👏 🌟 Jose Maria de las Rivas for taking the initiative to square up some data the team has been looking for a long time for! I appreciate your efforts!👏 🌟 Kyle Vidergauz for your insights into your Close customers. The support and guidance you give the team for "everything Close" is great! 👏 Let’s keep recognizing the people who make the work better and the wins bigger. It's so simple but it's what really matters daily! #Gratitude #Teamwork #CompanyCulture #EmployeeSpotlight #ThankYou