Sava Aims to Modernize $6.5 Trillion U.S. Trust Administration with Agentic AI

Sava, a new company backed by Y Combinator, is developing an "Agentic Trust Company" designed to modernize the administration of the estimated $6.5 trillion in U.S. trusts. The startup aims to streamline the historically complex and often inefficient process of managing trusts by leveraging artificial intelligence. Co-founder Nimit Maru, a two-time YC alumnus, initiated Sava after experiencing significant difficulties setting up trusts for his own family following the sale of his previous company.

Maru's personal experience highlighted the outdated nature of the trust industry, which often relies on manual processes, spreadsheets, and extensive paperwork. He noted the frustration of not being able to track the status of his multi-million-dollar trust, contrasting it with the ease of tracking a pizza delivery. This realization spurred him to return to Y Combinator's Fall 2025 batch to build a more efficient and transparent solution.

Sava's core innovation lies in treating the trust charter as programmable infrastructure, akin to how Stripe revolutionized payments for developers. The company's AI agents are designed to automate numerous administrative tasks, including data collection, signature chasing, document assembly, and reconciliation of positions. This approach ensures continuous compliance through automated risk checks and a timestamped audit trail for every step, reducing errors and increasing efficiency.

The platform is being developed to serve a broad range of clients, including fintech companies seeking to embed trust and estate services into their offerings, estate attorneys looking for faster client onboarding and live dashboards, and financial advisors needing real-time API access to client positions. Sava's approach allows these entities to innovate on client-facing experiences while the platform handles the complex regulatory and fiduciary "grunt work." The company is currently operating with its trust charter pending approval.