After an AI training programme, leadership asks one question: did it work? Most L&D teams answer with feedback scores. Those show that people enjoyed the session. They don't show whether anything changed.
The problem
AI training is usually measured at the wrong level. The Kirkpatrick model names four: reaction (did people find it useful), learning (did they gain the skill), behaviour (do they use it on the job) and results (did the organisation's outcomes move). Feedback forms stop at the first.
Three things make the higher levels hard to measure for AI training:
No baseline was taken, so there is nothing to compare against
The measures are vague ("productivity", "adoption") and nobody owns them
Measurement ends on the last day of training, before behaviour has had time to change
The solution
Pick a small number of measures tied to the team's real work, take them before training, and take them again 30 and 90 days after. Keep the set small enough that a team lead can collect it without a project.
The approach
Before training
Skill. A short practical task on the team's own material, scored against a rubric: brief quality, output accuracy, review catches.
Usage. How often each person uses AI, and for which tasks. A five-question survey is enough.
Work measures. Two or three numbers the team already tracks: turnaround time for a standard deliverable, review cycles per document, backlog size.
During training
Record the shared assets the team builds: prompt templates, review checklists, workflow maps. These are evidence of learning and the tools that keep behaviour going.
30 and 90 days after
Repeat the practical task and the usage survey
Pull the same work measures
Ask team leads one question: what does the team now do differently?
Reporting to leadership
One page: baseline, 30-day and 90-day figures for each measure, the assets the team now owns, and one example of changed work. Be honest about what didn't move. A measure that stayed flat tells you where the next intervention belongs.
Where we fit. Every AI Content Ops programme includes pre- and post-assessment and a 30-day check-in. Programme 03 adds a full 30-day implementation window. Book a discovery call.
