Six months ago your organisation ran an AI awareness session. Attendance was good and the feedback forms were positive. Today two or three people use AI regularly. Everyone else works the way they did before.
We hear this from L&D and HR leaders more than anything else. The session was usually fine. Awareness sessions were never designed to change daily work.
The problem
People are already using AI, mostly without help. Microsoft and LinkedIn's 2024 Work Trend Index surveyed 31,000 people in 31 countries. Three in four knowledge workers used generative AI at work, 78% of them brought their own tools, and only 39% had received any AI training from their employer.
A single session does little to close that gap, for three reasons.
Generic examples. Demonstrations use sample content: a made-up email, a generic summary. Back at their desks, people face their own release notes, course outlines and campaign briefs, and each person has to connect the demo to the work alone. Most don't.
No practice. Watching someone prompt well and prompting well are different skills. Forgetting starts fast: Ebbinghaus showed in 1885 that most forgetting happens within a day of learning, and a 2015 replication found the same curve.
No shared standard. The few who keep going each invent a private method. Outputs vary, reviewers don't know what to check, and managers conclude AI isn't reliable enough for real work.
The solution
Time spent practising matters. BCG's AI at Work 2025 survey of 10,635 employees found that 79% of people with more than five hours of training were regular users, against 67% of those with less. Leadership support moved the needle further: positive sentiment rose from 15% to 55% where it was strong.
A team that changes how it works needs four things:
Practice on its own documents, briefs and review processes
A shared way of working: templates, prompts and review checks everyone uses
Time inside the actual workflow, with the actual tools
A measure of where the team started and where it ended
The approach
Before you commission the next AI session, ask four questions.
Will people practise on their own work, or on examples?
Will the team leave with a shared method, or with individual tips?
What will you measure before and after?
Who supports the team in the weeks after the session?
If any answer is unclear, the next session will probably go the way the last one did.
An L&D manager from a Bengaluru GCC who attended our STC India programme put the difference in exactly these terms: pre- and post-assessment, a brief they could take to their manager, and outputs the team kept using after the sessions ended.
Where we fit. Reliable AI Workflows (Programme 02) is two days of hands-on practice on your team's own content, ending with a shared method the team keeps. Book a discovery call.
