Level 02 · Skilled

Comfortable Using AI

You already use AI tools most weeks. This track is about getting noticeably better output — not more tools, better technique.

Last updated: September 2026

What separates skilled use from casual use

Most people plateau at "type a question, accept the first answer." Skilled use means giving the model enough context to do the job right the first time, and knowing when to push back on its answer.

Habits worth building

Context-first prompting. Lead with the situation, the audience, and the constraint — not just the task.
Iterative refinement. Treat the first answer as a draft; ask for specific changes rather than starting over.
Cross-checking. For anything factual, verify against a second source before you use it.
Tool-matching. Use a research-first tool (like Perplexity) for facts, and a reasoning-first tool (like Claude) for structured writing or analysis.

Try this: take a prompt you'd normally send in one line, and add three sentences of context — audience, format, and one constraint. Compare the two outputs side by side.

Where to go from here

If you're starting to write code around AI models, or want to understand what's happening underneath the interface, move to the Technical track.

Go deeper with these lessons

Turn regular use into a repeatable method

At this level I would stop measuring progress by how many AI tools you have tried. A better test is whether you can get a useful result repeatedly. Pick two or three tasks you genuinely do — for example summarising research, drafting customer communication, or organising a plan — and improve the process around those.

A practical four-step workflow

Brief: state the job, audience, source material and constraints before asking for an answer.
Inspect: look for missing assumptions, vague claims and invented facts instead of judging only the writing style.
Refine: ask for targeted changes. “Make it better” teaches you very little; “remove repetition and preserve these three facts” does.
Verify: check important claims against primary or reliable independent sources before using the output.

Skilled checkpoint

Move on when you have at least one workflow where AI saves time without lowering the quality of the final work, and you know exactly which parts still require your own judgment.