Finpilot Secures $10M Series A to Build AI for Institutional Investors

Finpilot co-founder and CEO Lakshay Chauhan announced the round on November 4, 2026: a $10 million Series A led by Lobby Capital, with participation from Madrona, to build an AI-native operating system for institutional capital.

The company develops software for teams overseeing large, long-term pools of money at universities, public pensions, foundations, hospital systems and families. Finpilot names Seattle Children’s, Children’s Health of Dallas, Hackensack Meridian Health, TIFF, Blue Cross Blue Shield and the Wyoming Retirement System among the institutions using its software. Their investment work helps finance activities including medical research, health care and people’s retirements.

A Record Built for Investment Offices

The operational target is the recurring work required to make investment information usable. Investment data is spread across older systems that do not connect, while many reports arrive as PDFs by email. Staff must read those documents, extract figures, check them against other information and enter them in the right place each month and quarter. The resulting numbers feed board and committee decisions, making accuracy and clear sourcing essential.

Finpilot organizes its platform into three layers. The first maintains a current record of an institution’s holdings and documents. The second maps relationships between records, including what each number means and where it originated. The third consists of agents that read incoming reports, update figures, prepare analysis and reports, and track changes. When judgment is needed, the agents refer the issue to a person, and each number remains traceable to its source.

Users can describe the views they want and have the system build them on live data, then keep those views current. Knowledge about how a team prefers work done is saved in the system for the wider organization, rather than remaining in an individual’s chat history.

Agents Stop at Defined Review Points

The agent workflows are configured processes rather than one-off prompts. Agents monitor approved sources and prepare work against mandates, previous notes and other approved material. Defined review points stop the process when a person must decide how to proceed. In a quarterly-monitoring example, an agent receives a manager letter, drafts an update and escalates a material key-person change before the update advances.

Supported tasks span manager and portfolio monitoring, research briefs, investment-committee and board materials, client reporting, meeting preparation and summaries, and approved updates to customer relationship management (CRM) systems and other records. The meeting workflow can separate decisions, owners and unresolved questions from general discussion, then propose follow-ups and record changes that move forward only after approval.

The underlying Connected Data layer links existing CRM, research, portfolio, document and communications systems while leaving each system authoritative for the records it owns. Entity resolution connects different references to the same investment, organization, person or account and preserves their origins. Uncertain matches and conflicting values stay visible for review, while values retain their document, page and system provenance. Permissions determine which records each user can access.

Security Controls and Open Roles

Finpilot’s security documentation reports SOC 2 Type II assurance and encryption in transit and at rest. Each institution’s records, documents and retrieval index remain within a separate boundary, with authentication, roles and permissions checked on every request. Customer data is used to provide the contracted service and is not used for model training under the company’s approved policy and applicable provider contracts. Missing, conflicting or consequential information becomes a review item, keeping qualified staff responsible for judgment.

The company is also recruiting in Seattle. Its careers page lists Seattle-based openings for a software engineer, product manager, marketing lead, account executive, business development representative, AI solutions specialist in client success, and business operations associate.

The funding supports a platform that connects records, analysis and recurring workflows in one system. Its central challenge is to reduce manual preparation without obscuring sources or weakening accountability. The product’s design makes that trade-off explicit: agents prepare and advance the work, while people retain approval over consequential decisions.