Most people use AI at a fraction of its capability — one-shot questions, accept the first answer, move on — and never learn the handful of techniques that separate a frustrating tool from a force multiplier. The gap isn't a better model; it's how you use the one you have. This reads how you use AI now, picks the two or three techniques that would help you most next, and shows the difference on your own examples — a targeted level-up, not a firehose of a hundred tips.
What This Skill Produces
- An honest read of your current use — how you use AI now and the specific habits capping your results (one-shot asking, no context, accepting first drafts, no examples)
- Your highest-leverage next techniques — the 2–3 that would most improve your results, from: giving rich context, iterating instead of accepting, showing examples, decomposing big asks, assigning a role, and verifying output
- A before/after on your own use — the same task done your current way vs the leveled-up way, so the payoff is concrete
- A simple practice path — how to build the new techniques into habits one at a time, not all at once
- The mindset shift — from "search box" to "collaborator you brief, iterate with, and check" — the frame behind all the techniques
Required Inputs
Ask for these if not provided:
- How you use AI now — a real example of a prompt or task (shows exactly what to level up)
- Where it frustrates you — where results fall short (points at the missing technique)
- What you use it for — your main tasks and domains
- Your level — beginner / regular / trying to go advanced
Framework: Close The Gap One Technique At A Time
- Diagnose the current pattern. Look at how the person actually uses AI — the habit capping their results is usually obvious (one-shot, context-starved, first-draft-accepting) and points straight at the fix.
- Pick the 2–3 that pay off most. Don't dump every technique. From context-giving, iterating, showing examples, task decomposition, role-setting, and verifying — choose the few that fix this person's bottleneck.
- Show the payoff on their example. Take their real task and show it done the leveled-up way beside their current way — the gap is the motivation.
- Reframe the mindset. The through-line: AI is a collaborator you brief richly, push back on, and iterate with — not a vending machine you query once. The techniques follow from the frame.
- Practice one at a time. Build one technique into a habit before adding the next — layering beats a firehose that nothing sticks from.
Output Format
Leveling up your AI use
How you use it now: [current pattern + the habit capping results]. Add next (highest leverage for you): [the 2–3 techniques — e.g. give context, iterate, show examples]. Before → after (your task): [same task, current way vs leveled-up way]. The mindset shift: [search box → collaborator you brief, push, and verify]. Practice path: [one technique at a time → habit → add the next].
Quality Checks
- Diagnoses the person's actual current usage pattern
- Picks 2–3 targeted techniques, not a hundred tips
- Shows a concrete before/after on their own example
- Names the collaborator mindset behind the techniques
- Gives a one-at-a-time practice path
Anti-Patterns
- A firehose of tips with no prioritization.
- Generic advice not tied to how the person actually uses AI.
- Telling without showing the before/after payoff.
- Technique lists with no mindset shift underneath.
- Everything at once instead of one habit at a time.
Example Trigger Phrases
- "How do I get better at using AI?"
- "I feel like I'm using AI at 10% of what it can do."
- "How do power users actually use AI so well?"
- "Teach me to get dramatically better results from AI."
- "What am I doing wrong that makes my AI results so basic?"