Skip to content

Funding Brief|Kastle Raises $24M Series A to Build an AI Workforce for Banks

Sep 20, 20261 min
TL;DRKastle raised a $24M Series A led by Insight Partners, with its agents having processed more than $1.8 billion in transactions. The round signals that large financial institutions don't have to choose between living with a legacy system's limits and spending years replacing it — Kastle is betting that agents layered directly on top of existing systems are the fastest real path to enterprise AI adoption.

🌏 中文版

Funding Details

FieldValue
CompanyKastle (San Francisco, US)
RoundSeries A
Amount$24M
Lead investorInsight Partners
Follow-onY Combinator, Commerce Ventures (continuing existing investors), Fifth Wall (new), plus a group of financial-services executives joining as individual angels
ValuationNot disclosed
Total raisedNot fully disclosed (this Series A is $24M; the company previously raised a Y Combinator-backed seed round)
Founded2024 (Y Combinator S24 batch)
HeadcountAbout 10 (per Y Combinator's company page, pre-round)

What the company does

Kastle builds an "AI workforce" for financial services, starting with consumer lending — letting banks and lenders hand large volumes of high-frequency, time-sensitive work to agents without ripping out their existing core systems.

The product deploys specialized agents that work directly on top of a financial institution's existing core systems, executing high-volume workflows and keeping systems of record current in real time. That's a deliberate departure from "replace the legacy stack with an AI-native one," which typically means a multi-year, high-risk migration in a world where the underlying technology shifts every quarter. Kastle describes the result as "hybrid teams": agents handle high-volume, repeatable work while people focus on complex cases that require judgment, expertise, and relationships — and agent-completed work still runs inside the institution's existing processes and controls.

Co-founder and CEO Rishi Choudhary was previously the founding product manager for Redfin's Mortgage Marketplace, growing that business from $3M to $6M in revenue at age 22. Co-founder and CTO Nitish Poddar previously led engineering at Verkada, building a patented motion-detection system and the company's entire backend inference infrastructure. The two graduated from Y Combinator's S24 batch in 2024 and now serve several of the world's largest enterprises and banks. The press release's main body states agents have processed more than $1.8 billion in transactions, while the "About Kastle" boilerplate later in the same release says more than $2 billion — likely because the company-description template wasn't updated to match the latest figure, but both point the same direction: transaction volume is growing fast enough that even the press release can't keep its own numbers in sync.

What this round signals

What it means for the agent ecosystem

The new capital is earmarked for expanding engineering, product, and go-to-market teams; deepening platform capabilities; and accelerating deployments with banks and financial institutions across North America. Kastle is naming a structural "adoption gap": financial institutions know what AI can do, but they're constrained by the systems and operating models they already run — replacing the core is too slow and risky, while bolting a conversational AI assistant onto the old system only captures a fraction of AI's potential. Kastle is betting on a third path: agents layered on top of existing infrastructure, rather than a replacement for it.

What investors are betting on

Insight Partners Managing Director Rebecca Liu-Doyle put the logic directly: "Financial institutions do not need another layer of software that creates more work for their teams. They need AI that can reliably complete the work while ensuring compliance. Kastle deploys AI agents that can navigate complexity, pass the bar on regulatory rigor, and get high-stakes work done without waiting for a multiyear transformation." That points to a larger read: in a heavily regulated industry like financial services, whether an agent can pass regulatory scrutiny is itself the entry bar — so what Kastle is really selling isn't raw technical capability, it's the ability to execute reliably inside an existing regulatory framework.

Numbers worth watching

  • Agents have processed more than $1.8 billion in transactions (the press release's "About" section says more than $2 billion) — remarkable transaction volume for a company that only graduated from YC in 2024, approaching the scale of some established fintech players
  • A team of roughly 10 people is already serving "some of the largest enterprises and banks in the world," pointing to a business that can support high-value transactions on a very low headcount
  • Insight Partners manages over $90 billion in assets, has invested in more than 900 companies, and has seen 55-plus portfolio companies reach IPO — it's a heavyweight growth-stage software investor rather than an early-stage seed fund, so having it lead a Series A is a relatively uncommon pairing that suggests Insight sees Kastle as a company that can scale quickly

Watchlist status

Kastle is not yet tracked in the watchlist. Recommend adding it under section D6 (Financial Services / Insurance), alongside Hebbia, Hadrius, Taktile, and Gradient Labs, with tracking focus on: agents layered directly on top of banks' existing core systems to execute consumer-lending workflows without requiring a full system migration.

Today's takeaway

The usual bottleneck people cite for AI adoption in financial institutions is model accuracy, but this round is a reminder that for large regulated institutions, the real bottleneck is whether AI can be adopted without blowing up existing core systems and compliance processes. That's likely why Insight Partners — a fund that typically invests at the growth stage — was willing to lead a Series A for a company founded in 2024: the bet isn't on the underlying model technology, it's on whether "agents layered on top of legacy systems" becomes the standard way financial services sidesteps expensive system rebuilds.

References