Two popular fixes for stalled AI adoption: hand out a prompt pack, or run a workshop. Both feel like doing something. Both usually change very little three months later.
They fail for the same underlying reason, even though they look completely different.
The shared mistake
A prompt pack assumes the barrier to adoption is not knowing what to type. A workshop assumes the barrier is not understanding what the tool can do. Both are addressing information gaps.
Neither is addressing the actual barrier, which is habit. Most engineers who are not using AI tools well are not stuck because they lack the right prompt or missed the demo. They are stuck because they have never built a repeated, low-risk practice of reaching for the tool on a specific kind of task, under real deadline pressure, with real stakes if it goes wrong.
Information does not build habit. Repetition does.
Why prompt packs specifically fall short
A prompt pack is a static document, usually a list of example prompts for common tasks, sometimes with a bit of context on when to use each one. It gets shared once, referenced a few times in the first week, and then quietly forgotten because nobody's actual workflow includes "check the prompt pack" as a step.
Even when the prompts are good, they solve the wrong problem. The gap for most engineers is not "I don't know what to type." It is "I don't have a routine that includes trying this in the first place." A better prompt does not create that routine.
Why workshops fall short in the same way
A workshop delivers the same one-time information transfer, just live instead of written. It can be more engaging, answer questions in real time, and build more excitement in the room. But the moment it ends, engineers go back to their desks with the same lack of a real, repeated workflow they had before, just with better awareness of what is theoretically possible.
Watching someone else use a tool well does not automatically transfer into your own habits. That transfer requires doing it yourself, on your own work, more than once.
What breaks the pattern
The fix for both is the same, because the failure mode is the same: stop trying to transfer information faster, and start building repetition into how the team actually works.
That looks like:
- picking one or two specific, recurring tasks the team commits to doing with AI, not a broad menu of possible uses
- making that the default for those tasks for a defined stretch of time, not an optional experiment
- defining what good output looks like so people know if they are doing it right
- checking back in a few weeks to see who actually built the habit and who didn't
A prompt pack or a workshop can be a fine input to that process. Neither is a substitute for it.
What good looks like
If your team has a prompt pack sitting in a wiki nobody opens, or ran a well-attended workshop that did not move usage numbers afterward, that is not evidence AI adoption doesn't work here. It is evidence the fix stopped at information transfer instead of continuing to habit-building.
If you want help building a rollout that goes past the document or the session and actually changes how the team works, book a call. If I'm not the right person, I'll say so.