Sela Raises $21M as AI Agents Originate $1B in Mortgages Monthly

Mortgage technology startup Sela has raised $21 million in new funding, with the company reporting that its AI agents now help originate roughly $1 billion in mortgages every month. The raise signals investor confidence in AI-driven automation as a core engine for loan origination rather than a peripheral efficiency tool.

Sela's pitch centers on deploying AI agents to handle parts of the mortgage origination workflow — the kind of document-heavy, decision-intensive process that has traditionally required large teams of loan officers and processors. The funding will presumably support scaling that agent-based model further into lender operations.

Mortgage origination has long been a benchmark for how painful a regulated, paperwork-heavy process can feel for both customers and employees. If AI agents are now materially contributing to $1 billion in monthly originations, that is a signal the technology has moved past pilot status and into real operational load-bearing roles — not just chat-based assistance, but agents executing steps in a workflow that affects underwriting speed, cost, and borrower experience.

For leaders in digital transformation, this is a concrete data point in the broader shift from AI as a productivity add-on to AI as infrastructure embedded in core financial services processes. For experience leaders, the implication is that speed and consistency in lending — historically a source of borrower frustration — may become a competitive differentiator decided by which lenders adopt agentic AI fastest and most responsibly.

— $21 million raised by Sela in its latest funding round

— $1 billion in mortgages reportedly originated monthly with the help of Sela's AI agents

A headline funding number is easy to report; what's harder to assess is whether the underlying experience actually improves for the people taking out these mortgages. Automating origination can cut cycle times, but it can also strip out the human checkpoints borrowers rely on when a six-figure decision feels confusing or stressful.

The real test for Sela and any agentic AI player in lending isn't origination volume — it's whether borrowers can tell the difference between faster and better. Speed without transparency just moves the anxiety earlier in the process; it doesn't remove it. Any lender adopting this kind of automation should be measuring trust and comprehension alongside throughput, because a billion dollars a month in originations means nothing if borrowers can't explain why they were approved, denied, or priced the way they were.