Kastle, a San Francisco startup focused on artificial intelligence for financial operations, has closed a $24 million Series A round to expand its platform of specialized AI agents for consumer lending.
Insight Partners led the investment, with existing backers Y Combinator and Commerce Ventures returning and new participants including Fifth Wall plus a collection of founders and financial-services executives.
The company is positioning itself as a builder of an “AI workforce” for banks and lenders rather than another layer of chatbots or a full replacement for core banking systems.
Its agents are designed to sit on top of the technology institutions already run, completing high-volume, time-sensitive tasks such as payment collection, borrower inquiries, servicing updates, and related back-office work.
Kastle says those agents have handled more than $1.8 billion in transactions for some of the largest banks and enterprises, with later company materials citing a figure above $2 billion.
That integration-first approach is the heart of the pitch.
Large lenders know AI could reshape operations, but they are constrained by fragmented legacy platforms, manual processes, and strict regulatory requirements.
A complete shift to an AI-native stack can take years and carry substantial operational risk.
Simple assistants bolted onto existing tools, meanwhile, capture only a sliver of the potential.
Kastle argues it offers a third option: hybrid teams in which software agents take on repeatable work and keep systems of record current, while human staff concentrate on cases that need judgment, empathy, relationship management, and compliance oversight.
Consumer lending is the initial beachhead because the work is both high-volume and tightly controlled.
Origination, servicing, collections, and loss mitigation typically span multiple systems and must meet exacting service, risk, and regulatory standards.
Kastle’s agents are built with that domain context in mind so they can act inside approved workflows instead of merely routing tickets to a queue.
Co-founder and CEO Rishi Choudhary has framed the choice facing institutions as a false binary: live with the limits of current operations or endure a long, risky rip-and-replace of core systems.
“We give financial institutions an AI workforce that can operate across the systems they already have, so they can capture the benefits of AI now—not five years from now,” he said.
Insight Partners managing director Rebecca Liu-Doyle echoed the emphasis on reliability and compliance, arguing that lenders need AI that finishes work rather than adding more software for staff to manage.
The new capital will go toward growing engineering, product, and go-to-market teams, expanding platform capabilities, and speeding deployments with banks and lenders across North America.
The timing aligns with a broader industry shift from isolated pilots toward more systematic use of agentic AI.
Kastle, founded in 2024 and previously supported by Y Combinator, already claims production use at a number of major US mortgage servicers. If the model holds, the company is betting that the fastest path to AI-native operations is not a greenfield rebuild but software that can work inside the systems banks already trust.