Finpilot co-founder and CEO Lakshay Chauhan. (Finpilot Photo)
Seattle startup Finpilot , which uses AI agents to take the paperwork off the plates of people who invest money for hospitals, pension funds and foundations, has raised $10 million in a Series A round.
The round was led by Bay Area venture firm Lobby Capital , with participation from Madrona , which led Finpilot’s $4 million seed round in 2024. Finpilot, which has raised about $14.5 million since its 2023 founding, plans to use the new funding to expand its product and hire across the company.
Its customers include in-house investment teams at institutions and firms that manage money on their behalf. In the industry, they’re known as allocators, because they spread capital across outside fund managers. Much of their information arrives from fund managers as PDFs and spreadsheets, and someone has to read the reports, pull out the numbers and enter them into the right systems.
Finpilot pulls that information into one place and has its agents handle much of the work. The idea is to let investment teams spend their time on judgment and decisions, not on reading data from documents and typing it into a database, said Lakshay Chauhan , Finpilot’s co-founder and CEO, in an interview.
“Nobody wakes up like, ‘All right, I’m excited to do that,’” he said.
It’s a change of course for the company. When GeekWire covered its seed round , Finpilot was building a ChatGPT-style tool for analysts researching public companies. It shifted to allocators in mid-2024, Chauhan said, after concluding that AI added only marginal value in public markets, where Bloomberg and S&P Capital IQ had already organized most of the data.
“There’s no Bloomberg for what they do,” he said of allocators.
Customers include Seattle Children’s, Children’s Health in Dallas, Hackensack Meridian Health in New Jersey, Blue Cross Blue Shield of Wyoming, the Wyoming Retirement System, outsourced investment firm TIFF, and Seattle wealth management firm Laird Norton Wetherby. Some of those institutions hadn’t changed their core systems in 15 years, Chauhan said.
Finpilot has about 28 institutional customers in all, and revenue has grown tenfold since the start of the year, he said. It hasn’t lost a customer yet, he noted.
How it works: Finpilot pulls an institution’s data from its portfolio management system, document storage and other sources into a single record, matching each fund, holding and client to one identifier. Users ask questions in a chat window and set up agents to handle recurring work.
In an example Chauhan described, an agent could review updates from 35 private equity managers at the end of a quarter, figure out what changed and point an analyst to the three things that need attention. Agents also prepare briefings before meetings with fund managers, and they ask a person for approval when a step needs one.
Some customers run Finpilot alongside their existing software. Others have used it to replace older customer relationship management and research management systems, where investment teams log their dealings with fund managers.
Chauhan said Finpilot most often gets compared with general-purpose AI tools such as ChatGPT and Claude, including finance-focused versions like Claude for Financial Services, and with legacy CRM vendors.
Chauhan said the difference is accuracy. A 10-page report from a chatbot has to be checked number by number before it goes to a chief investment officer. In Finpilot, every number links back to the document it came from, so it can be checked.
In finance, he said, “you cannot be wrong.”
Finpilot runs its own in-house model for the high-volume job of reading financial documents, and uses models from Anthropic, OpenAI and Google for other tasks. It doesn’t train AI models on customer data and keeps each customer’s data separate, Chauhan said.
Founders: Chauhan grew up in India and came to the U.S. for a master’s degree in aerospace engineering. He interned at Boeing in Seattle, but when he graduated, he said, many full-time aerospace jobs required U.S. citizenship or permanent residency, and he was in the country on a visa.
He went to work for an engineering optimization software company and studied quantitative investing on the side. That led him to a paper by John Alberg , co-founder of Seattle investment firm Euclidean Technologies, which uses machine learning to pick stocks. Chauhan called Alberg repeatedly to propose working together, and Alberg eventually hired him. Chauhan later led AI at Euclidean.
Alberg is Finpilot’s co-founder and executive chairman but isn’t involved in day-to-day operations. His late father, Tom Alberg, co-founded Madrona.
Finpilot has about 14 employees, up from four in 2024. All are based in Seattle except one in Boise, Idaho. The company recently moved into the AGC Building on Lake Union after outgrowing its old office.
What’s next: Finpilot plans to fill six or seven jobs in the next few months, in product, marketing, client success, sales and business development, and about as many early next year.
Part of the money will go to implementation teams, Chauhan said. Their job is to sit down with each customer, learn how its investment team works and show it how to redo that work with AI agents.
