AI Guides › Career & Work

What Managers Actually Want From Someone Who Uses AI Well

By Nigel Guy · 2 min read

The assumption people make is that using AI visibly and often is itself impressive — the screenshots of clever prompts, the "I got this done in ten minutes" comment in standup. Most managers don't experience it that way. What they're actually tracking is whether the work in front of them is reliable, and whether they need to check it.

The Trust Checklist

1. You can explain any part of it without the tool. If a manager asks "why this number" or "why this approach" and the honest answer is "the model suggested it," that's a problem — you should be able to defend every choice as your own.
2. You checked the parts that matter before handing it over. Not everything needs the same scrutiny, but the load-bearing claims — numbers, commitments, anything another person will act on — need your own verification, not the model's confidence.
3. You disclose it when it's relevant, without making it the headline. "I used AI to draft this" is one sentence in context, not the framing for the whole conversation. Bury the tool, lead with the result.
4. The judgement calls are visibly yours. What to include, what to leave out, what the actual recommendation is — a manager wants to see you deciding, not relaying.
5. You know where it's not helping. Being able to say "I tried this and it wasn't right for the situation" reads as more competent than using it on everything by default.

Where visible AI use actually helps

Mentioning it briefly when it demonstrably sped up something the manager cares about — turnaround time on a deliverable, more thorough first-pass research — lands well precisely because it's understated. It's evidence you manage your own workflow well, offered once, not repeated.

What to skip

Skip narrating your AI use as the achievement itself — "I used AI to build this" is not, on its own, a pitch for anything; see "What Actually Gets You Promoted, And Where AI Fits Into It" for the fuller version of this point. Skip volunteering every tool and prompt you used unless asked; it reads as insecurity about the output rather than confidence in it.

Guardrails

• What a given manager values varies a great deal — some want disclosure by default, some don't care, and a few have real policy concerns about tool use. Read the room and the actual policy before assuming this pattern applies uniformly.
• This is a general pattern based on how trust and delegation typically work in a workplace, not a guarantee about any specific manager's preferences.
• If your organisation has an explicit AI use policy, that overrides anything general advice suggests — check it rather than assume.

All 751 AI guides · JulieMango plans from £17/mo