Answer extracted from the Masters of Scale podcast — listen to the full episode below.
Why do 90% of AI investments fail? Systems thinking and organizational feedback loops
Organizations fail at AI investment because they become self-referencing systems closed to external feedback—convinced they already have all the answers. MIT Sloan research reveals that 90% of AI investments fail with zero return on investment, largely because leaders have lost the humility to acknowledge what they don't know when facing emerging technologies and disruption.
How closed feedback loops destroy organizational learning
A healthy organization operates like a living system with permeable boundaries where quality data, insight, and feedback flow in and out freely. When those boundaries harden and leaders stop actively seeking external input, something dangerous happens: the system becomes self-validating. Leaders believe their own internal narrative, assume their strategies are sound, and miss critical signals from the market, their teams, and emerging realities.
This dynamic is not incidental to AI failures—it's often the root cause. As Brené Brown explains in the episode, the organizations that succeed are those where leaders maintain genuine intellectual humility about the limits of their own knowledge, especially during moments of rapid technological change.
"A healthy system requires permeable boundaries where good data and feedback can flow in and out freely. When boundaries close and feedback is not brought in consistently, organizations become self-referencing systems that believe they are great and know everything."
Brené Brown — Author and Researcher, University of Houston (Graduate College of Social Work). Over two decades, Brown has worked directly with C-suite leaders and their direct reports, researching the core competencies that separate thriving organizations from those that stagnate. Her latest work, Strong Ground, synthesizes decades of research into courageous leadership and the specific skills required to navigate disruption.
The courage to admit you don't know
The 90% failure rate for AI investments is not a technology problem—it's a leadership psychology problem. Leaders who invest heavily in AI often assume they understand its implications, its risks, and how it fits into their strategic landscape. They don't ask hard questions of people outside their echo chamber. They don't invite contradiction or seek perspectives that challenge their assumptions.
The antidote is a specific kind of courage: the willingness to publicly acknowledge that in times of disruption and emerging technology, you know very little. This is not weakness—it's the foundation of adaptive learning and strategic resilience.
The MIT Sloan data is unambiguous: organizations that treat AI investment as a closed technical problem fail. Those that treat it as a systems challenge—requiring new feedback mechanisms, intellectual humility, and willingness to adapt strategy based on what actually happens—have a far better chance of generating real value.
This applies far beyond AI. Any major investment—transformation, acquisition, new market entry—succeeds or fails based on the organization's capacity to remain curious, to solicit hard truth from the outside, and to update its beliefs when reality contradicts its assumptions.
See also
What is emotional granularity and why is it critical for leadership effectiveness?
Emotional granularity is the ability to have a robust emotional lexicon and accurately name what you're experiencing. Research shows adults in this country can identify only about 3 emotions, while approximately 85 to 90 emotions are important for leadership effectiveness and personal wellbeing.
What is the distinction between cognitive empathy and affective empathy in leadership?
Cognitive empathy is when someone shares a problem, you reflect back what you hear, and they feel seen, heard, and believed—this is the foundation for trust and psychological safety in organizations.
How should leaders identify and overcome their personal defense mechanisms when facing uncertainty?
Leaders should identify their go-to armor—the self-protective behavior they default to when afraid, such as perfectionism, micromanagement, or cynicism—and recognize it as a sign they are in uncertainty, risk, and exposure rather than a sign of weakness.
Key takeaways
- Closed organizational boundaries prevent the flow of critical external feedback, causing leaders to become self-referencing and out of touch with market reality.
- MIT Sloan research documents that 90% of AI investments fail because organizations treat them as isolated technical projects rather than systems challenges requiring adaptive learning.
- The counterweight to high-failure-rate investments is intellectual humility—the courage to acknowledge what you don't know when facing disruption and emerging technology.
- Permeable boundaries allow organizations to remain continuously attuned to signals, challenge their own assumptions, and adjust strategy based on real-world feedback.