How Bridgewater Turned 100 Million Documents Into an Edge - Sam Green & Aaron Linsky, Bridgewater

Bridgewater Associates built the Vault: more than 100 million documents from 135 external sources plus the firm's own memos and meetings, cleaned, permissioned and tagged to its internal ontology. Sam Green (team lead, Applied AI) and Aaron Linsky (head of product engineering) share three lessons: humans and agents need different things from search, the biggest wins come from one tool that serves both, and the real prize is agents that proactively bring investors what deserves their attention. Includes how a hand-built benchmark took precision at 10 from about 60% to over 90%. Chapters 00:00 Intro: three lessons from building the Vault 01:11 Bridgewater's 50-year research process 02:35 Why unstructured data needed the same rigor 03:25 What's in the Vault 04:14 Open web vs the Vault: a side-by-side 06:53 Search for humans vs search for agents 08:19 Benchmarking precision: "Vault search is bad" 09:32 Fast filtering at scale 11:35 Getting investors to change their habits 13:02 What the Vault enables: analysts, forecasts, signals 14:05 Proactive agents: the investor workbench 15:44 How the Vault compounds what the firm learns Sam Green, Bridgewater Associates Aaron Linsky, Bridgewater Associates Join us in SF for the second annual AIE CODE summit: https://aie.codes our last flagship summit of 2026. Applications live now!