Giving the Work Back to the People

In the AI Story

The principle violates every organizational expectation about leadership. People want leaders to absorb their problems and return solutions. Leaders are selected, trained, and rewarded for exactly this problem-absorption capacity. Giving the work back requires the leader to refuse the role the organization expects her to perform—not because she lacks answers (though she often does) but because providing answers would prevent the organization from developing the adaptive capacity the challenge requires. The refusal produces disappointment, frustration, and the perception that the leader is not doing her job. Heifetz is explicit: this is the cost of adaptive leadership, and the cost is unavoidable.

In practice, giving the work back means several specific moves. It means asking the question and then giving the question to the people: 'What does it mean to be an engineer when AI writes code?'—asked not rhetorically but as genuine work assigned to the engineers, whose answer will emerge from their experimentation and struggle. It means protecting the space for that work against the pressure to fill it with technical activities: no premature plans, no rushed timelines, no substitution of expert answers for the learning the people must do. It means holding the frame while the content shifts: stable team structures, predictable meeting rhythms, consistent leadership presence—the container that enables people to tolerate the uncertainty of not knowing who they are becoming.

The principle requires the leader to distinguish between abandonment and appropriate delegation. Giving the work back is not leaving people alone with an impossible challenge. It is creating a holding environment—relational trust, temporal space, structural stability—that contains the anxiety while enabling the experimentation. The leader is intensely present, but present in a different mode than the answer-giving mode organizational culture rewards. The mode is closer to therapeutic holding than managerial directing: the therapist does not solve the patient's problem but creates the relational space in which the patient can confront the problem herself, with support that does not become rescue.

Heifetz described the AI application directly: organizations need 'a leadership that's generating more leadership' cascading through the system. The adaptive challenge of AI does not have a single answer developable at the top; it has thousands of local answers, each specific to the people and the work. The backend team's answer will differ from the frontend team's; the senior engineers' will differ from the juniors'; the local contexts will produce local adaptations. This distributed emergence is precisely what centralized transformation roadmaps cannot produce, because the roadmap assumes the answer is known and needs only implementation. Giving the work back assumes the answer is unknown and can only be discovered by the people whose identities are being reshaped.

Origin

The concept emerged from Heifetz's observation that in successful adaptive change, the authority figure's primary contribution was not the solution but the conditions enabling others to find solutions. He formalized this in Leadership Without Easy Answers (1994) and developed its operational implications through The Practice of Adaptive Leadership (2009), which provided diagnostic questions: Is this work the leader must do, or work the people must do? Is the leader solving to reduce her own anxiety or to enable others' learning?

The principle has been tested across thousands of organizational cases in business, government, and civil society. The failures are instructive: leaders who could not tolerate being seen as unhelpful, who provided premature answers under pressure, who took the work back when the organization's distress exceeded their comfort. The successes are quieter: organizations where adaptive work proceeded because leaders held the frame, protected the space, and sustained their own discomfort long enough for others' learning to produce genuine transformation.

Key Ideas

Violates the contract. Organizations expect leaders to solve problems on their behalf—giving the work back breaks this implicit bargain, producing disappointment the leader must tolerate because meeting expectations would prevent adaptive learning.

Not abandonment. Giving the work back requires intense leader presence in a different mode—creating conditions (holding environment, regulated distress, protected time) rather than providing solutions.

The people have the answer. Adaptive challenges can only be solved by those whose identities are being reshaped—no external authority can change someone's sense of self on her behalf; the transformation must be self-directed.

Distributed emergence. AI's adaptive challenge has thousands of local answers specific to contexts; centralized plans that provide uniform answers foreclose the learning from which context-appropriate adaptations emerge.

Leader holds the frame. While content (roles, contributions, identities) shifts, the leader maintains structural stability (teams, rhythms, presence)—the container that enables people to tolerate the uncertainty of transformation.