A note before the moves. This is written for the person on the receiving end of a mandate rather than the person issuing one, and it assumes the mandate is staying. Company policies vary, so where yours specifies tools or rules, the policy wins over anything here.
The announcement usually comes at an all-hands, delivered with the particular energy of leadership that has recently attended a conference. The company is now AI-first, licences have been purchased, and there may well be a dashboard tracking adoption.
Floating somewhere above all of it is an unspecified expectation that you will now do your job with a tool nobody has explained, to a standard nobody has defined, while producing exactly as much work as before. If your private reaction landed somewhere between an eye-roll and a low-grade dread, you are in the majority, and the research backs the dread: a majority of workers report AI sometimes costing them more time than it saves, which is what happens when a tool arrives before a reason.
None of that changes the situation, though, so here is the version of it that works, in three moves.
The short answer. A mandate survives best when you stop treating it as a demand to use AI for everything and start treating it as a demand you can satisfy on your own terms. Find the two tasks in your week the tool genuinely carries, route around its weak spots by using the mandated tool where it is actually strong, and make your wins quietly visible. That combination satisfies the dashboard, protects your time, and, more often than the eye-roll predicts, ends up handing you back a few hours.
Why mandates land badly, briefly
The failure is almost never the workforce. Mandates arrive as tool-first decisions, licence purchased and use case pending, which inverts the order any sensible adoption runs in.
The tool then gets tried on whatever task happens to be in front of each person, which for a good number of tasks makes things slower, and the slowness reads as personal failure until quiet resentment sets in. Naming that mechanism matters because it relocates the problem: you are not bad at AI, the mandate is bad at tasks, and the fix is choosing the tasks yourself before the mandate chooses them for you.
The three moves
| The move | What it looks like |
|---|---|
| Find your two tasks | Pick the two jobs in your actual week that the tools genuinely carry, usually the long document to summarise, the recurring drafting, or the notes that need structuring. Two is enough, and everything else can stay manual with a clear conscience |
| Route around the weak spots | Use the mandated tool where it is strong rather than everywhere. If it is Copilot, that means minutes, summaries of internal material, inbox drafts and formula help, rather than free-form writing, where it tends to disappoint |
| Make the wins visible | Mention the reclaimed hour in the standup, or share the summary that saved the meeting. Mandates measure attitude as much as output, and two visible wins buy you the standing to keep the rest of your workflow yours |
If the mandated tool in question is Copilot, the strengths and the gaps are worth reading properly before you decide which two tasks to hand over, and you have Copilot at work, do you still need ChatGPT sets out the split in more detail.
The move most people miss
There is a fourth move available once the first three are running, and it is the one that converts a mandate from an imposition into an asset. The mandated tool came with a boundary: it is sanctioned, which makes it the one place confidential work material can go without the anxiety consumer tools carry.
That inverts the usual complaint rather neatly. The stiff corporate tool everyone resents is the only one cleared for the sensitive documents, and the nimble consumer tools everyone prefers are the ones that need the caution. Run both accordingly, with the sanctioned tool for anything touching real work data and a consumer tool on your own time for general thinking with nothing identifying in it, and you have a two-tool setup most enthusiasts never bother to reason out.
The posture that wins
Loud resistance to a mandate spends credibility on a fight that is already settled, and performative enthusiasm burns your evenings feeding a dashboard. The posture that wins is quiet competence: two genuine uses, visibly working, everything else unchanged, and a reputation as someone who adapted without fuss.
Whatever comes out of the tool still needs a look before it leaves your screen, which is what the pre-send check is for. Mandates fade into furniture within a couple of quarters, and when this one does, the record will show you used the tools where they worked. That is all fluency ever was, and you will have got there without attending a single webinar.
Not sure where to start?
Your two tasks are findable in sixty seconds.
The free finder tells you which AI to open for the task in front of you, in about 60 seconds, and the full "AI, sorted." reference lands in your inbox.
Try the free finder →One question in, one answer out.
Clair helps non-technical professionals know when to trust their AI, when to check it, and when to skip it.