Council Post: The Next Enterprise AI Advantage Is Institutional Memory

AI in enterprise is often pitched as labor-saving automation, but GRAIL CEO and co-founder Mayank Kejriwal argues the bigger constraint is weak institutional memory. He describes fragmented knowledge—stored across inboxes, PDFs, spreadsheets, Slack threads, legacy systems, and a few experts’ heads—as an expensive bottleneck that slows execution and makes decisions less consistent. Kejriwal cites McKinsey’s estimate that knowledge workers spend significant time writing, searching, and collaborating, with productivity gains of 20% to 25% from better knowledge sharing. He also points to Deloitte’s warning that retirements could trigger a major knowledge exodus unless organizations capture critical know-how before experienced staff leave.