Organizations are complex systems composed of people who come and go, much like biological organisms are made up of individual cells that die and regenerate. But unlike biological systems, where each cell carries the organism’s full genetic code, institutional knowledge in organizations is fragmented: No single individual has the complete picture. For this reason, organizations invent mechanisms to capture, retain, and transmit knowledge, ensuring continuity, even as individuals cycle in and out. Lew Platt, former CEO of Hewlett-Packard, was right to lament: “If only HP knew what HP knows, we’d be three times more productive.”

It is no surprise then that the “knowledge management” market is projected to reach $2.5 trillion by 2030. This explosion in prospective value is due, in part, to the potential and progress of generative AI as a tool to excavate buried data and tacit knowledge, recontextualize it, and offer organizations not just a better understanding of their past, but a more coherent strategic vision of their future.

Today’s standard approaches to knowledge management in organizations are constrained by two limitations. First, they often fail to capture what may end up mattering most: The vast repositories of tacit knowledge that individuals seldom translate into neat, “final” documents. The reasoning behind key decisions, the context of near-misses and past actions, and the instincts that shaped an institution’s culture—these are the kinds of insights that quietly disappear, eroding an organization’s ability to learn from itself. Second, traditional systems are built on an assumption that organizations can anticipate what knowledge will be needed or useful in the future.

A recent artistic project led by our co-author, Zoé Vayssières, with Harvard University’s Digital Data Design Institute on the use of GenAI to uncover and inscribe Harvard’s early “memories” (based only on raw data, like salary records, not existing historical accounts) provides a powerful illustration of how technology can not only resurface institutional intelligence, but recontextualize it—making memory dynamic, discoverable, and usable in unpredictable, varying contexts.

This shift from static knowledge management to dynamic “memory management” unlocks new possibilities for complex organizations. Rather than relying on predefined categories, GenAI enables organizations to classify information at the point of retrieval, surfacing relevant insight based on the questions being asked, not the assumptions made at the time of storage. That responsiveness makes memory not just easier to access, but more actionable. This is not simply a technological shift, but a strategic one: For the first time, organizations can treat memory not just as something to preserve, but a resource to activate—turning accumulated experience into competitive advantage.