Blend Labs (NYSE: BLND) says its agentic AI platform for home lending, Blend Autopilot, has reached its tenth lender customer, marking a new adoption milestone six months after entering live-loan production and three months after commercial availability. The platform is processing an average of more than 3,500 new loans per week, according to the company, with AI agents supporting tasks including document review, income verification, pre-underwriting and closing quality control. The expansion highlights growing interest in automating operational workflows across mortgage lending, home equity and other financial services.

Blend Autopilot Reaches 10 Lenders

Blend Labs has signed its tenth lender to Blend Autopilot, an agentic intelligence layer designed to automate parts of the home lending process. The milestone comes as banks, credit unions and mortgage providers explore how AI agents can take on repetitive loan-processing tasks while keeping lending decisions and operational controls within established workflows.

Blend says Autopilot is now working on an average of more than 3,500 new loans each week, equivalent to an annualized pace of nearly 200,000 loans. The company introduced the platform into live-loan production six months ago and made it commercially available three months ago.

The latest lenders to sign on include Whatcom Educational Credit Union (WECU), a Washington-based credit union, and Gate City Bank. Blend says customers span community financial institutions, regional banks and mortgage providers, reflecting interest across different types of lending organizations.

The adoption milestone offers an early indication of commercial demand, although the number of signed lenders alone does not establish how broadly the platform has been deployed within each institution. Actual adoption will depend on production volumes, integration requirements, operational reliability and the ability to demonstrate measurable benefits.

AI Agents Target Manual Lending Work

Mortgage origination involves a sequence of information-intensive activities, including collecting borrower documents, reviewing income details, identifying missing information, preparing files for underwriting and checking documentation before closing. Delays can occur when documents are incomplete, manual reviews accumulate or borrowers must be contacted repeatedly to resolve outstanding requirements.

Blend Autopilot is designed to address parts of this process through purpose-built AI agents. According to the company, its agents can review loan files, identify required next steps and take action within lender workflows. Current use cases include document review, income review, borrower follow-ups, pre-underwriting and closing quality control.

The platform operates across mortgage and home equity lending. Rather than positioning AI as a standalone assistant, Blend describes Autopilot as an intelligence layer connected to its wider lending platform, which includes data, workflow capabilities and more than 150 integrations.

This architecture is intended to let lenders deploy specialized agents across different stages of the lending lifecycle. Blend also says customers can build their own agents through its Model Context Protocol (MCP) server, alongside using the company’s prebuilt tools.

For financial institutions, this approach could help reduce repetitive administrative work and improve the consistency of loan-file processing. However, lenders still need controls for data access, borrower privacy, error handling, auditability and human escalation when automated systems encounter exceptions.

Reported Operational Improvements

Blend cites earlier production results showing a 10% to 15% increase in loan pull-through, reductions in cycle times of two to four days, and an average of 4.5 hours of fulfillment work automated per loan.

These figures are company-reported results, and the announcement does not provide the underlying sample size, comparison methodology or details about how outcomes vary by lender. They should therefore be treated as reported performance indicators rather than independently verified industry benchmarks.

If sustained, shorter processing times and less manual fulfillment work could benefit lenders by helping staff handle application workloads more efficiently. Borrowers may also experience fewer delays when supplying documents or responding to requests, particularly when they submit information outside normal business hours.

The platform’s potential value depends on more than document-processing speed. Mortgage lenders must also maintain appropriate underwriting standards, ensure that decisions comply with applicable requirements and retain accountability for actions taken on a borrower’s application.

Onity Mortgage Highlights After-Hours Processing

Onity Mortgage, which Blend describes as serving 1.4 million customers, is among Autopilot’s early adopters. The lender began with the platform’s pre-underwriting agent, according to the announcement.

Patrick Joyce, Onity Mortgage’s vice president of digital products, said the technology supports a more responsive self-service experience. He described how borrowers can upload documents during evenings or weekends and have the agent review them automatically, rather than waiting for staff to resume work. If a document is unsuitable, the system can prompt the borrower to provide the correct information.

This example illustrates a practical use case for agentic AI: moving routine document checks closer to the point at which information enters the lending process. Earlier identification of missing or incorrect documents can reduce unnecessary handoffs between borrowers and lender staff.

The extent of the benefit will depend on how accurately the agent identifies document requirements and how well the system handles unusual cases. Lenders must also ensure that automated borrower communications are clear, appropriate and consistent with their policies.

Building AI Across the Lending Lifecycle

Blend’s broader strategy is to connect intelligence across customer engagement, application orchestration, document processing, loan officer productivity and closing quality control. The company says Autopilot uses the same underlying lending data, integrations, workflows and controls developed through nearly 15 years of platform operations.

A connected approach could help financial institutions avoid deploying isolated AI tools for individual tasks that do not share context or workflow state. It may also allow lenders to introduce automation incrementally, starting with document review before extending to additional stages.

For Blend, the next test is whether early customer adoption translates into sustained production usage and repeatable operational improvements across a wider range of lenders. For banks and mortgage providers, the key considerations will be measurable efficiency, implementation complexity, security, regulatory compliance and the quality of human oversight.

Blend Autopilot’s tenth-lender milestone marks progress in the commercialization of agentic AI for home lending. Its long-term significance will depend on whether the technology can deliver consistent results while preserving the controls and accountability required in a high-stakes financial process.

Market Landscape

Mortgage and home equity lending involve document-heavy workflows, repeated borrower communications and multiple operational handoffs. Financial institutions are exploring AI agents to automate routine tasks, improve processing speed and help staff manage growing workloads.

Blend’s approach embeds purpose-built agents within a connected lending platform rather than treating automation as a separate application. Its reported results suggest potential improvements in cycle times and fulfillment effort, although independent validation and further deployment data would be needed to assess the broader impact.

Key adoption considerations include integration with existing systems, data security, model accuracy, audit trails and human oversight. Automation must support lender processes without weakening underwriting standards or regulatory accountability.

Top Insights

• Customer milestone: Blend says its Autopilot platform has reached 10 lender customers, three months after commercial availability.
• Loan processing volume: The platform is working on more than 3,500 new loans per week, according to company-reported figures.
• Workflow automation: AI agents support document and income review, borrower follow-ups, pre-underwriting and closing quality control.
• Reported performance: Blend cites 10%–15% higher pull-through, two to four days shorter cycle times and 4.5 hours of fulfillment work automated per loan.
• Connected AI infrastructure: Blend offers purpose-built agents and an MCP server for customers seeking to build their own agents on its lending platform.

Get in touch with our fintech expert