Why reassurance backfires
A statement like "don't worry, nothing will change" only works if it's true – and in AI rollouts, it's almost never fully true. As soon as the first concrete changes become visible, the blanket assurance is exposed as either a lie or leadership's own ignorance. That costs more trust than the original uncertainty ever would have.
What transparency concretely means
Transparent communication explicitly distinguishes between three categories: what's already settled, what's actively being decided, and what's still genuinely open. So instead of "nothing will change", something like: "These three tasks will very likely change, these two are still being decided, and for these areas there's currently no reliable assessment yet." That's less comfortable than a smooth assurance, but it holds up to later scrutiny.
Handling the task-risk question
Concretely, that also means: if there's a real risk in certain task types or entry-level positions, it should be named openly – along with the equally honest point that this risk isn't distributed evenly across the whole workforce. That distinction is harder to communicate than a blanket yes or no, but it's the only statement that matches the actual facts.
Practice section: the three-category template
Before any major communication about an AI rollout, a simple template with exactly these three columns pays off: already settled, still being decided, currently open. Every statement in the communication can be assigned to one of these three categories – and any statement that can't be assigned to any of them is probably reassurance you should cut.