Most people quit AI within 3 weeks — and it's usually your best people. Microsoft tracked 300,000 workers using AI: excitement spiked for about three weeks, then most quietly gave up. Not the laziest people — the most careful, switched-on ones, the exact people you most want using it. The fix isn't a better tool. It's realizing AI isn't a tool skill, it's a management skill. Treat it like a sharp but green new hire, and these six skills are how you manage it. None of them are technical — you already have every one. You've just never aimed them at AI.

How to use this playbook

Each skill below comes with one drop-in prompt. Copy it, fill in the [brackets], and paste it into Claude, ChatGPT, or any AI. Run them in order on your next real task and watch the output stop feeling like "AI slop" and start feeling like a sharp employee did it.

1. Context — brief it like it's day one

Vague ask in, vague answer out. Your best new hire would need a real brief before touching the work — so does AI. Stop dumping a one-line request and start handing over the goal, the audience, and what "good" looks like.

📋 The Day-One Brief prompt:

You are my new [role] on day one. Before you do anything, here's your brief:
- What we're trying to achieve: [the real goal / outcome]
- Who it's for: [audience / customer]
- What "good" looks like: [an example, or the standard to hit]
- Constraints: [tone, format, length, deadline, what to avoid]
- What I've already tried: [context so you don't repeat it]

Restate the task back to me in your own words, list anything you're missing, then do it.

2. Judgment — know when to trust it and when to check

AI says everything with the same confident voice, whether it's right or guessing. Your job isn't to trust it or distrust it — it's to know which parts to verify before you act.

📋 The Trust-or-Verify prompt:

After your answer, add a "Confidence Check":
- Rate your confidence 1-10 and say why.
- List the 3 things in this answer most likely to be wrong or made up.
- Tell me exactly what I should verify myself before I rely on this.
Never present a guess as a fact — label anything you're unsure about.

3. Break it down — hand it tasks it can actually handle

Give AI a giant, fuzzy task and you get a giant, fuzzy mess. Great managers slice big work into pieces a new hire can nail one at a time. Make the AI do the slicing first.

📋 The Decomposition prompt:

Don't solve this all at once. First, break [big task] into the smallest sensible sequence of steps you can each do well.
Show me the plan as a numbered list, then stop and wait for my "go" before starting step 1.
We'll work through them one at a time.