Kastle, a San Francisco startup that builds AI agents to handle customer service and collections for mortgage lenders, announced September 17 it raised $24 million in a Series A round led by Insight Partners. Y Combinator, Commerce Ventures, Fifth Wall and a group of unnamed founders and financial services executives also participated. The company declined to disclose its valuation.
The capital arrives as Kastle, which graduated from Y Combinator's Summer 2024 batch, has processed more than $2 billion in transactions across 10 of the top 25 US mortgage servicers, according to its announcement. That figure offers a glimpse of traction in an industry historically slow to adopt new technology but now under pressure to automate labor-intensive operations.
Kastle's core product is what it calls KastleOS, a platform that deploys autonomous AI agents across lenders' existing loan-servicing systems. The agents handle voice calls, chat, email and SMS without requiring infrastructure overhauls. They're trained on a thicket of consumer-lending regulations: FDCPA, TILA, RESPA and TCPA, among others. Compliance, in this business, is not optional.
Co-founder and CEO Rishi Choudhary spent time as a founding product manager for Redfin Mortgage Marketplace and later led the company's search experience team. His co-founder, Nitish Poddar, worked as a tech lead at Verkada, where he built motion search and backend inference infrastructure. Both studied at the University of Illinois Urbana-Champaign.
The company has signed at least two marquee customers. Carrington Mortgage Services selected Kastle in July as its enterprise AI agent partner, deploying the platform across contact center and borrower-facing servicing operations. "Kastle stood out because they are specialized in the mortgage servicing space, and have received strong positive feedback from clients already using their solution," Elizabeth Balce, executive vice president of loan servicing at Carrington, said in a July statement.
New American Funding announced a partnership the following month to deploy Kastle's voice AI for end-to-end loan interactions and round-the-clock self-service. "Kastle AI enables us to serve each customer with greater speed and consistency throughout the life of their loan," Christy Bunce, president of New American Funding, told HousingWire at the time.

In April, Kastle announced a direct integration with ICE Mortgage Technology's MSP servicing system, one of the industry's most widely used platforms. The integration allows Kastle's AI agents to retrieve and act on data inside MSP without manual handoffs.
Kastle claims its platform delivers 70% lower cost per call, 40% lower handle time and customer satisfaction scores above 90%. The company did not provide third-party validation of those metrics, which is common for early-stage startups pitching operational efficiency gains.
The competitive landscape is filling out quickly. Fiserv launched agentOS, an agentic AI operating system for banking workflows, in May. Abrigo released APX, an agentic platform for community bank lending, in July and claimed reductions in manual work north of 40%. ICE Mortgage Technology has also been adding AI call prediction and voice-agent capabilities directly into MSP, which could present both a partnership opportunity and a competitive threat for Kastle.

Fifth Wall, the real-estate-focused venture firm that joined the round, framed the investment around compliance. "Voice is the last major unautomated service layer in mortgage," the firm wrote in a LinkedIn post. "What sets Rishi Choudhary and his team apart is not simply their voice technology. It is the compliance infrastructure underneath it."
That positioning matters. Financial institutions have been flooded with AI pitches over the past two years, many of them half-baked or ill-suited to regulated industries. Kastle's pitch is different, perhaps more pragmatic: work within existing systems rather than replace them.
"Kastle creates a third path," Choudhary said in the announcement. "We give financial institutions an AI workforce that can operate across the systems they already have."
Rebecca Liu-Doyle, managing director at Insight Partners, echoed the practical angle. "Financial institutions do not need another layer of software that creates more work for their teams," she said. "They need AI that can reliably complete the work while ensuring compliance."
Kastle plans to use the capital to expand engineering, product and go-to-market teams and to accelerate deployments in North America. The company did not specify hiring targets or provide a timeline for expansion into other verticals, though mortgage servicing is vast enough to support years of growth if the technology holds up.
