AI Training & Workshops
When AI adoption fades after the first week
A practical review of task fit, access, instruction quality, and ownership when a team stops using a new workflow.
← The Signal — BlogIn this piece
If a team stops using an AI workflow after an enthusiastic introduction, investigate the work before scheduling another demonstration. People may lack access, dislike the output, spend too long checking it, or simply not encounter the task often enough.
“Week two” is a convenient point to check in, not a universal adoption deadline. This revised article replaces an unverified client anecdote with a diagnostic exercise a manager can use.
Ask someone to show the complete task
Watch the person find the inputs, use the tool, inspect the result, and put the output where it belongs. Include the steps outside the assistant. Copying information between systems may account for more effort than generating the draft.
Ask which part they would change first. A request for better source material may be more informative than a request for another prompt.
Distinguish four possible problems
| Problem | Evidence to look for | Next action |
|---|---|---|
| Access | Participants cannot use the required account or information | Resolve permissions with the owner |
| Task fit | The method is slower or harder than the existing approach | Narrow the task or keep the original method |
| Output quality | Similar omissions or unsupported claims recur | Review sources, instructions, and evaluation examples |
| Ownership | Nobody maintains the method or answers questions | Name an owner and a feedback route |
This is a recommended diagnostic, not a validated scoring model. More than one problem may apply.
Make a small correction and observe it
For an illustrative meeting-summary task, the team might find that the assistant assigns owners where none were agreed. Add an explicit requirement to preserve missing owners, then test it against new notes and the earlier failure.
Have participants compare the revised output with the source. A better prompt does not remove the review responsibility.
Decide whether to continue
Record total effort, necessary corrections, and whether the final artifact meets the team's standard. If the method is useful, keep practicing before adding another task. If the process needs different software or cleaner information, make that a separate decision.
Usage counts can show whether people opened a tool. They cannot by themselves tell you whether the work improved. Preserve room for staff to report that an experiment was not useful.
For a fuller approach, read Building practical AI skills across a team. AI Training covers tailored sessions and optional follow-up; ongoing support depends on the agreed scope.
Put this into practice
Bring the decision you’re working through.
A working session is a chance to review your goal, the tools or website you have, and the help you need. We’ll identify a sensible next step and discuss scope before any project commitment.
AI Training for Teams →Book a working session