Blend Labs Inc. is opening its lending platform to third-party AI agents, launching a new “Autopilot MCP Server” that allows financial institutions to build and deploy their own AI-driven workflows directly inside the origination process.
The company said the new capability is built on the Model Context Protocol (MCP), an emerging open standard that enables AI agents to securely connect with enterprise software and execute tasks across systems.
With the release, Blend is positioning its platform as an execution layer for AI, allowing lender-built agents to operate within existing loan workflows, data environments, and compliance frameworks.
“Autopilot MCP Server lets lenders bring their own AI agents into Blend and have them take action directly within loan workflows,” said Nima Ghamsari, co-founder and CEO of Blend Labs Inc. “This is a major step toward a future where AI doesn’t just assist — it actually does the work.”
What It Does
Autopilot MCP enables AI agents to carry out core lending tasks such as pulling credit, checking pricing, and verifying compliance, according to the company.
Those agents operate using each lender’s specific rules, overlays, and product guidelines, rather than a one-size-fits-all automation model.
The system also includes controls and audit trails for all agent activity, with sensitive actions, including rate locks, credit pulls, and disclosures, requiring lender authorization before execution.
Why It Matters For LOs
The move points to a next phase of AI adoption, one that goes beyond embedded tools and into customizable, institution-specific automation.
Blend’s earlier Autopilot release automated document review, compliance checks, and borrower follow-ups in real time. That positioned AI as a tool to accelerate workflows without replacing underwriting decisions.
The MCP server expands that approach, shifting from task-level automation to AI agents that can act across the full loan lifecycle, from application intake to underwriting support.
That could begin to change how work is divided between LOs and technology. Routine steps like condition clearing, data validation, and compliance checks may increasingly be handled by agents, allowing originators to focus more on borrower interaction and deal structuring.
From Efficiency To Execution
The shift builds on lenders' existing use of Blend’s platform to increase production capacity.
In a recent deployment, Northfield Savings Bank implemented Blend across mortgage and home equity lending to streamline workflows and enable LOs to handle more volume without adding staff.
Autopilot MCP extends that model, moving from faster workflows to AI agents that can execute tasks directly within them.
The launch also signals a broader competitive shift among mortgage tech providers.
By adopting MCP, Blend is positioning its platform as an open ecosystem rather than a closed LOS-style environment.
In practical terms, that means lenders are no longer limited to vendor-built automation. Instead, they can develop or integrate their own AI agents tailored to their workflows, products, and risk tolerances.
