Playbook 4 — Building Ability
The coaching layer between trained and competent
Ability is earned through repetition with feedback, not a single session. The week after training is where the knowing-doing gap opens, and it is the week most rollouts stop paying attention.
ADKAR puts ability after knowledge for a reason. Knowing how to do something and being able to do it under real conditions — deadline, impatient client, imperfect data — are separate states. Training gets someone to knowledge. Nothing about the training itself gets them to ability.
What gets them there is a coaching layer, and the useful part is that it does not need to be elaborate. Three ingredients tend to do the work.
Champions first. Pick one or two people per team who took to the tool naturally — usually whoever was already using AI on their own account — and make their role explicit. Not a new job title, just a named person colleagues can ask without it becoming a formal escalation. The best support is lateral, because asking a peer is cheaper socially than asking a manager.
Office hours second. A fixed, short window once a week where anyone can bring a real task and get unstuck. The value is not the hour itself; it is that problems surface instead of festering. A person who hits a wall and gets help in ten minutes stays a user. The same person who hits a wall alone at 6pm goes back to the old way.
Practice with feedback third. The ability stage needs repetition on real work, reviewed by someone who can tell the difference between a decent output and a plausible one. Without that feedback, people optimize for whatever looks finished, which is how bad AI output becomes normal.
A cadence that works:
- Named champions per team, not a central helpdesk.
- A weekly office hour, same slot, low ceremony.
- Each user has practiced on their own recurring tasks at least twice.
- Someone reviews early outputs and gives real feedback, not just approval.
- You are tracking who is still asking for help versus who has gone quiet — because going quiet is ambiguous, and often means stopping.
The instinct at this stage is to answer a dip in usage with another training session. That rarely works. The dip is a support problem, and training is not support.
Next: Playbook 5 covers what happens when the launch energy fades.