When the tide goes out, you see who’s swimming naked.
In the AI era, the tide is reversing course. Global AI spending will top $200 billion this year, yet most companies get disappointing returns. Leaders blame the tech. Bad data. Immature models.
They’re wrong. The real problem is human. It’s not what people can’t learn. It’s what they refuse to unlearn.
Executives spend 93% of AI budgets on technology and just 7% on people. That’s an expensive error. By 2030, nearly 40% of today’s core skills will be obsolete. Upskilling is essential, but it’s only half the job. The other half is strategic forgetting.
Organizations that fail to unlearn old mindsets are nearly three times more likely to miss critical market shifts. Learning is addition. Unlearning is subtraction. In a world drowning in information, subtraction is the more valuable skill.
This isn’t about memory loss. It’s about courageous abandonment. You clear the mental clutter so new ideas can grow.
What Unlearning Really Is
Unlearning isn’t amnesia. It’s the conscious choice to drop knowledge, processes, and assumptions that once worked but now hold you back.
It happens on three levels:
- Individuals: The analyst who must stop manually cleaning data because the AI does it faster. Their new job is interpretation.
- Teams: The product group that abandons a waterfall process that ensured quality for a decade but now kills speed.
- Organizations: The enterprise that drops the assumption that “zero risk” is the only acceptable stance.
Here’s what most leaders miss: unlearning isn’t the opposite of learning. It’s a precondition for it. You can’t pour new wine into old wineskins. Unless you evict the old tenants, the new ones can’t move in.
Why Unlearning Hurts So Much
If learning is downhill, unlearning is a steep climb. Why?
1. Past success is a curse
We repeat what works. The more successful a habit, the harder it sticks. As Deloitte’s China Widener put it: “We’re not asking people to give up things that failed. We’re asking them to give up things they’ve done for 20 years that worked.” It’s easier to abandon a failure than a formula that made you rich.
2. Stress makes us retreat
McKinsey’s research shows a cruel paradox. When leaders need fresh thinking most, stress defaults them to the most familiar patterns. They cling to the known. Accumulated knowledge becomes an anchor, not a sail.
3. Power gets in the way
Old processes come with old power structures. A VP who built their career on a legacy framework isn’t just being asked to change a spreadsheet. They’re being asked to shrink their own kingdom. Unlearning is political, and politics resists change.
4. It drains brainpower
Most change programs ignore the “interpretation gap.” When you introduce AI, you leave employees alone to figure out what it means for their jobs. This cognitive tax exhausts them. We’re asking people to run a marathon while adjusting their shoes every mile.
Frameworks That Work
Theory helps. Here are three tools to guide the mess.
Govindarajan’s Three-Box Solution
- Box 1: Manage the present.
- Box 2: Selectively forget the past. (This is the one we skip.)
- Box 3: Create the future.
Leaders obsess over Boxes 1 and 3. Without rigorous focus on Box 2, the past hoards resources and starves the future.
The STARS Method
Researchers at the University of Sydney offer a path:
- Scout: Find the hidden assumptions driving bad decisions.
- Trace: Figure out why you started doing this.
- Appreciate: Acknowledge why it worked back then (this builds empathy).
- Rebuild: Co-create new mental models with your team.
- Storytell: Replace the old narrative with one that celebrates letting go.
Unlearning as a cycle
It’s not a one-off event. It’s three overlapping steps:
- Awareness: Realize the old way is broken.
- Abandonment: Stop doing it.
- Relearning: Embed the new way until it sticks.
You’ll run this cycle again and again. Knowledge expires faster than ever.
Who Got It Right (and Who Didn’t)
Failure: Borders Bookstore
Amazon didn’t kill Borders. Borders killed Borders. Its management clung to the superstore identity. They believed physical footprint was an unshakeable advantage. They spent years in denial, then tried to unlearn too late. Their ability to “weather the storm” was precisely the problem—it prevented them from pivoting early.
Failure: AI Training Without Permission
A tech CEO mandated a full quarter of AI training. It consumed 20% of payroll hours. Yet 80% of employees refused to adopt the tools. Why? They didn’t lack skills. They lacked permission. Training teaches technique. Only unlearning changes a belief. The CEO never gave them permission to abandon the old manual processes that made them feel safe.
Success: The Firefighter Who Stopped Fighting
A high-growth CEO was the ultimate bottleneck. He’d spent 15 years as the “firefighter”—the guy who solved the unsolvable. To scale, he had to unlearn a core belief: “My value is measured by the problems I solve.” He replaced it with: “My value is measured by the problems my team solves without me.” Within six months, decision velocity jumped 40%. Retention improved. People felt truly empowered.
Success: The Banker Who Embraced Failure
A senior VP had 25 years in risk compliance. She was handed a fintech project. She applied her 200-page policy manual. The team stalled. She had to unlearn the belief that “absolute safety is the only path.” She shifted to “calculated failure.” Instead of asking “What could go wrong?” she asked “What’s the smallest experiment we can run to learn?” Her unit launched three successful pilot apps in record time. She became the firm’s new governance hero.
A Leader’s Playbook for Letting Go
How do you make unlearning a routine? Follow these five steps.
1. Audit your assumptions
You can’t abandon what you can’t see. Run “Assumption Audits” for every major business line. Ask: “If we started this market today, would we structure ourselves this way?” Most of your practices are relics of a dead environment.
2. Give explicit permission
People need a clear signal that the old rules are suspended. Use a town hall, a new charter, or a physical token. Shift the conversation from “solving” to “exploring.” Hold the space. Don’t jump in to fix the anxiety.
3. Make sense-making a ritual
Don’t leave people alone to decode change. Start a weekly “What’s Expired?” meeting. Ask three questions:
- What changed in our world this week?
- Which internal practice no longer fits?
- What will we stop doing?
4. Admit your own obsolescence
This is non-negotiable. If you want your team to unlearn, you must publicly admit when your old strategy is dead. If the CEO won’t abandon their pet projects, no one else will abandon theirs.
5. Rewrite the story
Organizations run on stories. To kill an old practice, you must stop telling the heroic story of its origin. Start telling the heroic story of its retirement. Celebrate the VP who shrinks her own department because AI does the work. Create a “Hall of Fame for Forgotten Things.”
Conclusion: Master Subtraction
We live in a world obsessed with adding. More data. More algorithms. More skills. But the winners will be those who master subtraction.
Strategic unlearning isn’t soft. It’s a hard competitive edge. Companies that practice it see revenue growth 32% higher than peers who just pile new initiatives on top of old ones. They’re lighter. Faster. Less burdened by gravity.
Learning feels good. It’s acquisition. Unlearning feels painful. It’s loss. But pain is the price of growth.
Next time you launch an AI project, don’t just ask: “What do we need to learn?”
Ask the harder question: *“What do we need to abandon?” *
The answer will decide if you transform—or become a footnote.
Sidebar: The Unlearning Check
Give yourself a raw score on these five:
- Can you name three processes your team uses that predate your current CEO?
- Has your team formally stopped doing anything in the last six months?
- Do your incentives reward new behaviors—or old outcomes?
- When did you last celebrate retiring a failed practice?
- If you cut 50% of internal reports, would decisions actually get worse?
For deeper work, revisit Govindarajan’s “Three-Box Solution,” Edgar Schein’s culture research, and the STARS framework.**For further reading, explore McKinsey’s work on the “Three-Box Solution,” Edgar Schein’s organizational culture research, and the STARS framework for mental model transformation.