Why Leadership Training Stops Working in Real Conversations

Companies are spending more on leadership development than ever. Mordor Intelligence projects the corporate leadership training market will reach $39.73 billion in 2026, growing to $57.64 billion by 2031. Mid-level managers account for the largest share of that demand because they sit at the center of execution and organizational change. The money is real. The question is whether the capability is.

The evidence suggests it often is not. INSEAD Knowledge reports that global spending on leadership development is estimated at around $100 billion a year, and that McKinsey found only 11% of executives strongly agreed their leadership interventions achieved and sustained the desired results. Michael Beer and colleagues at Harvard described the pattern as the "great training robbery". The capabilities built in classrooms rarely translate into behavioural change. The programme ends. The old environment resumes. The new behaviour evaporates.

The gap is not knowledge. It is application under pressure. The HR Digest reports on the Predictive Index's 2026 survey of 399 managers and 208 CEOs, which found that 74% of CEOs and business leaders believed their managers were "very confident" handling difficult workplace conversations. Only 20% of managers preferred to prepare for those conversations without a framework or guidance. And 42% said they knew what they wanted to say but struggled with how to say it.

The leaders thought the training had worked. The managers knew it had not. That is the moment when capable people freeze in the room.

Gregg Hannah has spent years working with that moment. He is a Toronto-based leadership educator who has taught leadership theory for years, and he works with managers through TL Solutions. His view is that the problem is structural. Companies invest in models. Models are theory. Theory is useful, but it is not the same as a framework a person can use when a conversation turns.

"The models aren't giving them the framework that they can use," Hannah says. "You need to convert the theory into something usable. Leadership is both theory and action."

The freezing happens because real conversations do not follow scripts. A manager trained on a model expects a certain response. The employee says something the model did not anticipate. The manager has no rehearsed words for that moment, and the conversation stalls. Hannah describes the real world as "messy, clumsy, a big ball of yarn." The skill that matters is not recalled. It is adaptive communication in real time.

"I hate sales scripts because sales scripts are exactly that," Hannah says. "Somebody else wrote them. They don't mean anything to me. And when I use them with you, they don't mean anything to you either." The point is not that preparation is worthless. It is that preparation without the capacity to adapt is a script by another name.

Hannah's alternative is a framework rather than a script. The manager learns to hear what is being said, understand what is actually meant, and respond in the moment. That is harder than memorising a model, and it is the only thing that works when the conversation goes somewhere the training did not cover. "If you get this understanding," he says, "then you can begin to understand how to better interact with that person using the language that they need and want to hear."

Artificial intelligence has entered this space, and Hannah is cautious about it. AI can help a manager prepare, brainstorm approaches, or surface considerations they might have missed. It cannot replicate the human judgment required in the room. "It's a tool," he says. "It doesn't replicate human thinking. It doesn't replicate human emotions." The risk is that managers rely on it too much, treating a generated response as a substitute for their own analysis. The machine can be confident and wrong at the same time. A human being on the other side of the conversation will notice.

The practical implication is not that companies should stop investing in leadership development. It is that the investment needs to reach the moment that matters. A manager who understands a model but cannot use it when an employee pushes back has not been trained for the job they are doing. The question for any leadership programme is not whether participants leave knowing more. It is whether they leave able to do something different in the next difficult conversation. If the training ends and the behaviour does not change, what was the $100 billion for?