After a four-month preview involving more than 25,500 loans, the company's Autopilot platform is now commercially available for mortgage pre-underwriting
Mortgage technology provider Blend has moved its agentic artificial intelligence platform from pilot to production, announcing that five lenders have signed on as the first commercial users of Blend Autopilot, an AI-powered pre-underwriting assistant designed to automate document review, income calculations, and borrower follow-up during the mortgage application process.
Among the first adopters is Onity Mortgage, one of the nation's largest mortgage servicers, which expanded its partnership with Blend after participating in a four-month early preview.
According to Blend, participating lenders used Autopilot across more than 25,500 mortgage loans during the preview period, providing ongoing feedback that helped shape weekly product updates before the platform became commercially available.
Unlike many recent AI announcements centered on demonstrations or limited pilots, Blend said Autopilot has already been tested on live mortgage files and is now available to lenders using the company's mortgage platform.
"Autopilot is live, working on real loans, and improving every week," said Nima Ghamsari, co-founder and head of Blend. "Waiting for a mortgage loan to close can be one of the most stressful parts of buying a home. Agentic AI can help lenders complete more work earlier, reduce delays, and give borrowers faster answers. That's the future of lending we're building with our customers."
Autopilot reviews borrower documents as they are uploaded, compares them against lending guidelines, identifies missing or inconsistent information, calculates GSE qualifying income, and generates contextual requests for additional documentation before a file reaches underwriting.
Blend said the goal is to address loan conditions while borrowers remain actively engaged rather than discovering issues days later after momentum has slowed.
The company said loans processed with Autopilot during pre-underwriting moved faster and converted at higher rates than comparable loans using traditional workflows, although it did not disclose specific performance metrics.
Onity Expands AI Strategy
Onity Mortgage said it has already seen the platform accelerate condition management by identifying missing borrower information and resolving documentation issues earlier in the loan process.
"Onity Mortgage has seen consistent impact across all condition types, surfacing and clearing conditions faster, resulting in quicker response times," said Colin Friday, senior vice president of consumer lending at Onity Mortgage.
The lender also plans to expand its use of Blend's AI technology beyond pre-underwriting through Blend's MCP Server, which allows lender-developed AI applications to securely access loan data and platform capabilities.
According to Onity, that next phase will include AI-powered voice agents capable of answering routine borrower questions around the clock while giving customer service representatives immediate access to loan information during live conversations.
"When a customer calls, our people should not have to search across multiple systems for an answer," Friday said. "The AI assistant can surface it for our support staff in real time so they can focus on the conversation and the customer experience."
The rollout follows Blend's continued expansion beyond traditional mortgage point-of-sale technology. Earlier this year, the company broadened its partnership with MeridianLink, extending its digital origination capabilities across mortgage, home equity and consumer lending. The addition of Autopilot builds on that strategy by bringing AI-driven workflow automation deeper into the loan manufacturing process.
What It Means
For originators, the significance isn't that AI is replacing loan officers — it's that lenders are beginning to automate repetitive tasks that often consume valuable time before a file reaches underwriting.
Autopilot is now commercially available to Blend mortgage customers, with the company saying future enhancements will continue to be driven by customer feedback, production performance and ongoing evaluation.