Across industries, leadership teams are accelerating AI adoption, launching pilots, and pushing toward scale. The challenge is not ambition; it is traction. Many organizations are doing the right things on paper and still seeing adoption stall once programs move beyond initial enthusiasm.
What often gets diagnosed as a readiness problem is something else entirely. AI introduces new capabilities, but it also forces people to weigh what they might gain against what they feel they are giving up. That calculation is rarely rational, and it is almost never neutral.
In practice, AI adoption is less about introducing new tools and more about understanding the psychological cost of letting go of what already exists.
This is where the concept of the endowment effect, first demonstrated by Kahneman, Knetsch, and Thaler, becomes critical (1). In their well-known experiments, participants who owned an object demanded significantly more to give it up than others were willing to pay to acquire it, often by a factor of roughly 2.5x. Ownership alone inflated perceived value, even for something as simple as a mug.
This dynamic remains highly relevant today. As highlighted in Berger’s The Catalyst, endowment is a core barrier to change: people are inherently attached to what they already have, and that attachment increases the perceived cost of switching (2).
AI strategies rarely account for this explicitly. That is the gap.
Endowment Is Already Making Everyday Technology Change Hard
Before introducing AI, most organizations are already navigating endowment in traditional technology shifts.
Consider a CRM migration. Even when the new system is objectively better, adoption often slows under familiar patterns of resistance. Teams defend the current system not because it is optimal, but because it is known. They understand its quirks. They have built workarounds. They know how to succeed within it.
From a purely functional perspective, switching tools requires process change. From a behavioral perspective, it requires giving something up. That “something” includes familiarity, confidence, and a sense of control.
Key takeaway: Even non-AI technology transitions are not just operational changes. They are perceived loss events, and the endowment effect amplifies that loss.
AI Introduces a New Layer of Endowment
When organizations move from non‑AI processes to AI-enabled ways of working, the nature of the transition changes.
Traditional system changes ask people to learn a new interface. AI asks something more personal. It asks people to rethink how they generate value in their role.
What individuals perceive themselves as “owning” is no longer limited to a tool or workflow. It includes the skills, judgment, and patterns of thinking that define how they operate day to day. The endowment is attached to capability.
Because of this, the perceived loss is greater. Replacing a system affects how work is done. Introducing AI affects how people see themselves within that work.
Key takeaway: AI adoption does not only replace processes. It challenges perceived ownership of expertise, increasing the psychological barrier to change.
The Overlooked Transition: Consumer AI to Enterprise AI
A second, often underestimated transition is already underway in many organizations. Employees are not starting from zero with AI. They are beginning with tools they have already adopted independently.
When organizations introduce enterprise-approved AI tools, they are not simply enabling new capability. They are asking people to switch from tools and workflows they already trust.
By that point, users may have developed their own prompts, habits, and ways of interacting with AI. These become part of how they work. In behavioral terms, they become endowed.
From an organizational perspective, moving to enterprise-grade AI improves governance, security, and scalability. From an individual perspective, it can feel like giving up something that already works.
Key takeaway: “Shadow AI” is not only a governance issue. It is an endowment issue. Users are being asked to trade something they already value, which raises the perceived cost of adoption.
Why Traditional Adoption Approaches Fall Short
When AI adoption stalls, organizations typically focus on improving training, communication, and enablement. These are necessary, but they assume the core barrier is knowledge or awareness.
The endowment effect suggests a different diagnosis. The issue is not that people do not understand the benefits. It is that the benefits are being weighed against an inflated perception of what is being lost.
Usage metrics, while useful, do not capture this dynamic. They show activity, but they do not show the internal calculation individuals are making about whether change is worth it.
This is why strong business cases alone rarely drive sustained adoption. They address the upside without addressing the perceived cost of leaving the current state.
Key takeaway: Adoption efforts that focus only on enablement overlook the primary barrier. The perceived loss of the current way of working is often stronger than the perceived benefit of the new one.
Reframing AI Change Planning
Once endowment is treated as a predictable force, the implications for AI strategy are immediate.
First, organizations need to explicitly account for what people feel they are being asked to give up. This includes not only tools, but also autonomy, familiarity, and competence.
Second, change narratives need to shift from replacement to amplification. When AI is framed as replacing existing work, it activates defensive behavior. When it is framed as building on existing strengths, it reduces the perceived loss.
Third, value realization approaches should expand beyond usage to include perceived trade-offs. Understanding where users feel friction or loss provides a more accurate view of adoption risk than activity metrics alone.
Key takeaway: Effective AI adoption planning requires designing for perceived loss, not just promoting potential gain.
AI Strategy Requires a Different Starting Point
Taken together, these dynamics point to a simple but often overlooked truth.
AI adoption challenges are not primarily technical, and they are not primarily informational. They are behavioral.
Endowment has always been present in technology change. What makes AI different is the scale of what is being displaced. The shift is not just from one system to another, but from one way of working to another, and in many cases, from one definition of individual contribution to another.
Organizations that recognize this early have an advantage. They design change strategies around how people actually respond to loss, rather than assuming a compelling vision of the future will be sufficient.
Key takeaway: AI adoption accelerates when organizations treat resistance not as a lack of readiness, but as a predictable response to perceived loss.
____________________________________________
Sources:
- Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1990). Experimental tests of the endowment effect and the Coase theorem. Journal of Political Economy, 98(6), 1325–1348. https://doi.org/10.1086/261737
- Berger, J. (2020). The catalyst: How to change anyone’s mind. Simon & Schuster.
