Leadership Anxiety

Most transformation post-mortems land on the same suspects: weak strategy, poor execution, resistant employees, not enough urgency. A new piece in the Fall 2026 issue of MIT Sloan Management Review argues we've been looking in the wrong place.

In "How Leadership Anxiety Derails Transformation", INSEAD's Declan Fitzsimons, Gianpiero Petriglieri, and Jennifer Petriglieri draw on a four-year ethnographic study of a professional services firm: roughly 760 hours of observation and more than 300 interviews across the full life of a change effort. The firm had everything the textbooks ask for: a thoughtful CEO, a committed team, a plausible vision, a participative process. It still failed, and it was eventually acquired.

Their diagnosis is that the problem was unacknowledged anxiety, starting at the top. For anyone leading in AI right now, that deserves a hard look.

The four stages of "defensive organizing"

The authors describe a pattern they call defensive organizing. In this pattern the organization's energy is quietly redirected from changing things to keeping leaders' anxiety at bay. It unfolds in four stages:

1. The comforting vision. Performance dips and leaders feel it personally. Instead of naming that worry, they adopt a reassuring story about the future. In the firm they studied, the story was "collaboration." The warning sign is that nobody asks for evidence that the idea will work. There's a lot of planning and very little real change.
2. The convenient foe. When the vision doesn't deliver, leaders look for a reason outside themselves. Project teams whose findings point back at leadership get attacked as the messengers. Resentment builds, and people stop caring.
3. The empty ritual. Engagement meetings, "small wins" videos and communication campaigns multiply. People act committed in public and are cynical in private. The authors argue that this is less "change fatigue" than change initiative fatigue: what drains people is the dissonance, not the change.
4. The reckoning. The organization weakens, and blame lands on a scapegoat, often the leader whose vision once inspired everyone.

Their most provocative line, and the one I keep returning to, is that the real purpose of many a strategic vision is to give leaders a story about the future that makes the present feel manageable.

Why this hits the AI industry especially hard

When the authors list the kinds of "we can do something" moves leaders make under pressure, one is a large investment in artificial intelligence.

That should make all of us in this field uncomfortable. Right now "AI transformation" is the ultimate comforting vision. It's plausible, it's urgent, and it affirms leadership. Questioning it feels naive. That combination is exactly what lets idealization take hold.

I see the stages play out in AI programs all the time:

• Comforting vision: "We're becoming an AI-first company." Everyone nods, and nobody asks what will actually change in how decisions get made, who owns what, or how leaders themselves will work differently.
• Convenient foe: The pilot stalls, and it becomes the data team's fault, or legal's, or "middle management's resistance." The finding that leaders never gave up control of the workflows the AI was supposed to reshape goes unheard.
• Empty ritual: Adoption dashboards, prompt-a-thons, mandatory usage targets. Usage climbs and value doesn't. In the corridors people say it's theater.

This connects to last week's theme on permanent AI disruption. Some of the exhaustion we blame on the pace of AI may actually be the cost of the gap between what leaders say about AI and what people experience.

The antidote: courage, curiosity, and care

The authors offer three capacities, each countering one stage of the pattern. Here's how I'd apply them to AI leadership.

Brave the worry. Anxiety can't be controlled, but it can be managed, and the first step is naming it. Before presenting your AI roadmap, try opening with what worries you: "I'm not sure yet where this creates real value for us, and I suspect some of you feel the same." Saying this isn't weakness. It keeps the team from rushing into reckless action because everyone is too anxious to think.

Find your truth tellers. Identify the people close to the actual work who will tell you whether your AI vision holds up. Ask them two questions: What would have to change in how I work for this to succeed? and What would have to change in how the business works? When you feel contempt for a group that "just doesn't get AI," the authors suggest asking yourself what that judgment is protecting you from.

Care visibly. Stop trying to close the gap between public enthusiasm and private doubt by pushing for more commitment. Ask whose efforts are going unrecognized and who is pretending to be fine. Then ask what is working and what isn't, and cut what isn't. That distinction matters in AI, where the pressure to keep launching new pilots is constant.

Anxiety in the face of uncertainty isn't a leadership failure; it's a normal part of leading. The failure is pretending it isn't there, because that's when it starts to shape your strategy for you.

Three questions to ask about your own AI transformation this week:

• Is anyone asking for evidence? If your AI vision met no challenge when you presented it, treat the silence as a warning sign rather than as alignment.
• Where is the blame going? If the explanation for slow progress always sits one level below you, look again.
• What are the corridors saying? The distance between your town hall and the hallway is a direct measure of how much defensive organizing is going on.

Your AI strategy should confront the organization's problems, not calm your nerves about them. The leaders who get the most out of this technology may not be the most confident ones, but the ones honest enough to say they're worried too and curious enough to find out why.

Inspired by "How Leadership Anxiety Derails Transformation" by Declan Fitzsimons, Gianpiero Petriglieri, and Jennifer Petriglieri, MIT Sloan Management Review, Fall 2026. The AI commentary and applications are my own.