Briefing
The situation
You have reviewed how your team spends its time. The picture is clear to you: a significant share goes to tasks an AI could handle — summaries, first drafts, reporting, reconciliations, manual data work. At the same time, these are the tasks the team has built its professional pride around. One person is known across the company for always having the numbers in order. Another uses the "unnecessary" reconciliations to catch problems early — but that isn't written down in any process description. You want to move the time to work that creates more value. You haven't yet said exactly what that work is. Here is what you don't know: whether the tasks carry something that doesn't show up in their output, whether there really is more valuable work to move the time to — or just an expectation of it, how the team will read the offer (as development, or as a first step towards becoming redundant), and who can actually judge quality once AI has done the job.
Discussion
Questions to wrestle with
What are we removing?
- 1.Which of these tasks carry something beyond their own output — relationships, quality control, learning, breathing room?
- 2.Who defined what counts as "value-creating" here — and was the team part of that definition?
What are we moving the time to?
- 1.Can we point to the new work concretely today? If not — what do we say to whoever asks?
- 2.What happens to the person whose whole professional identity sits in the task we are automating?
Framework · Four decisions per task
To lean on
Remove
The task never needed doing. Nobody asks for the output.
Keep deliberately
The task carries something beyond its output. We keep doing it ourselves, and we say why.
Assist
AI drafts, the human owns the judgement and the signature.
Automate
AI does it, the human spot-checks. Requires that someone can still judge quality.
Decision
Possible paths
- ALet the team map and classify its own tasks using the framework.
- BLeadership names three tasks to automate, with a firm deadline.
- CRun a time-boxed trial in part of the team and evaluate it openly.
- DDefine the new, more valuable work first — move nothing until it exists.
Triggers
Drop in when the discussion stalls
- ▸The most experienced team member says it straight out: "So you don't want me here."
- ▸After three weeks the AI does the same task in a quarter of the time — with one error nobody caught until the client called.
- ▸The team automates faster than you asked for, and starts using the freed-up time for its own priorities.
- ▸Someone asks whether the freed-up time means they can go home earlier.
For the facilitator
Tips to get more out of it
- Ask each participant to name a task in their own work that deserves decision 1 or 4. It changes the tone in the room immediately.
- If the group gets stuck on technology: ask what they would say, on the spot, to an employee in tears. That's the situation the change actually lands in.
- Option B is fastest and creates the most resistance. Option A is the most legitimate and risks nothing happening. Have the group acknowledge the trade-off rather than dodge it.
- Hold on to the question of what the time is for. If nobody can answer, the change isn't ready to be communicated.
Reflection
To take with you
- "If an employee describes this change to a friend a year from now — what do we want them to say?"