It’s been estimated that the cost of “corporate amnesia” — or inefficient knowledge sharing — can top tens of millions annually. Yet almost no one in the boardroom is talking about it.

AI may transform how organizations operate, but without a record of how they have made decisions, navigated crises, and earned trust over time, even the most sophisticated systems risk becoming disconnected from the experience that makes intelligence meaningful.

When CEOs Leave, the Lessons Leave With Them 

Leadership transitions are a natural part of organizational life. Boards often seek new leadership when markets shift or strategies change, and today’s environment — defined by technological disruption, geopolitical instability, and rapidly evolving consumer expectations — has only accelerated that cycle. Consider Boeing, Nike, and Stellantis: each navigated a recent CEO transition while simultaneously managing deep operational or reputational crises that demanded intimate knowledge of how the organization had failed and recovered before.

When CEOs depart, they rarely leave alone. Senior teams move on, long-tenured executives retire, and the institutional knowledge those leaders carry quietly disappears.

A company’s most valuable knowledge is often embedded in years or decades of decisions, pivots, failures, and breakthroughs. It lives in boardroom debates, cultural inflection points, product launches, regulatory battles, and moments when leaders had to choose between competing priorities under pressure. Without deliberate efforts to capture and preserve that experience, succession planning becomes little more than a leadership handoff — not a transfer of organizational knowledge.

Your AI Is Only as Smart as Your History 

This matters even more in the age of AI. As companies adopt generative AI and agentic systems, many assume that access to powerful models will create an advantage. In reality, the opposite may be true. If every company has access to similar AI systems, competitive differentiation may increasingly depend on the quality of the context from which those systems draw.

That context comes from experience — which, over time, becomes institutional memory: the accumulated record of how an organization has navigated complexity, balanced risk and opportunity, responded to crises, and built relationships with its stakeholders. Without that, intelligence — human or artificial — becomes generic.

Large language models can generate remarkably fluent answers, but without being grounded in a company’s specific history, decisions, and operating norms, those outputs often feel shallow or disconnected from reality. They can produce information. They cannot produce insight.

Just as the digital era required investment in data infrastructure, the age of AI will require infrastructure that preserves and activates institutional memory — including historical records, internal documentation, oral histories, and digital knowledge systems that allow organizations to learn from their own experiences. Increasingly, these systems will also inform AI tools themselves, grounding machine-generated insights in an enterprise’s real history rather than generic training data.

How Booz Allen Turned 110 Years Into a Strategic Edge 

When Booz Allen turned 110, CEO Horacio Rozanski didn’t just commission a retrospective. Through executive interviews, storytelling initiatives, and a digital archives platform, the company captured pivotal moments across its history — advising the U.S. Navy before World War II, supporting NASA during the space race — and connected those stories to its current identity as a technology and analytics leader. The effort wasn’t nostalgic. It was operational. 

The Archive Isn’t Missing — It’s Just Scattered 

The deeper problem isn’t preserving the memories of departing executives. It’s preserving the accumulated experience of the institution itself. That experience, often built over decades, lives in strategy documents, research reports, correspondence, internal publications, photographs, and other records that explain how the organization became what it is.

In many companies, this material is scattered across fragmented digital systems or buried in analog archives that have never been systematically organized. In the age of AI, that presents a strategic vulnerability. AI learns from data. If the knowledge that defines an organization’s experience is inaccessible or lost entirely, those systems will produce incomplete, generic intelligence. The question for leaders is no longer simply whether they will adopt AI — it’s whether they will deploy it in a way that reflects their organization’s hard-won experience.

Target illustrates what’s at stake. The company faces a leadership transition amid declining sales, brand confusion, and growing political pressure. Analysts argue that it has drifted from the distinctive identity that once set it apart — its design-driven merchandising, curated product mix, and cultural positioning that inspired shoppers to jokingly pronounce the brand’s name like a French boutique. What’s less discussed is whether any of that accumulated brand intelligence — the decisions, trade-offs, and creative instincts that built “Tar-zhay” — was ever formally captured. Or whether it simply walked out the door with the leaders who built it.

The challenge for an incoming CEO is not just operational — it is interpretive. Which aspects of the organization’s identity should be preserved? Which strategies have worked historically? Which lessons from past crises still matter?

Before your company deploys its next AI system, ask this: What does that system actually know about how your organization makes decisions? If the answer is “not much,” you haven’t built an intelligent enterprise. You’ve built a very fast amnesiac.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.