Why in-house AI champions burn out (and how to prevent it)

Nearly every AI initiative I've seen actually stick inside an organization has one identifiable person driving it: chasing stakeholder buy-in, translating between technical vendors and business teams, and personally absorbing the friction every time the new system doesn't behave the way someone expected. That role is rarely an official job title, and it's almost never resourced like the load-bearing function it actually is.

The burnout pattern is predictable: the champion is given the initiative as an addition to their existing job, not a replacement for part of it, and the personal credibility they've staked on the project's success becomes an ongoing source of stress every time adoption stalls or a stakeholder complains, with no formal authority to actually fix the underlying causes.

The organizations that don't lose their champion treat the role as a real allocation of time, not a volunteer position: a defined percentage of the person's week protected for the initiative, a clear escalation path so they're not personally absorbing every piece of organizational friction, and recognition, ideally including compensation, that acknowledges what the role actually requires.

When the champion burns out or leaves, the initiative rarely survives them, however good the underlying technology was. Protecting that person's capacity is not a nice-to-have around the edges of an AI project, it's closer to the actual load-bearing infrastructure the project depends on.