Creative range

Using AI beyond drafting — simulation, stress-testing, and divergent thinking.

Most people use AI as a typist. The operators who get outsized value use it as a cast of characters: a skeptical executive to rehearse against, a red team to attack the plan, a hundred-idea generator to break a creative block, a translator between formats no template covers. Creative range is a skill, not a personality trait — it has learnable moves.

1. Stage 1 · E1 FOUNDATIONAL

   Ask for many, then choose

   The first creative move is volume: instead of asking for the best tagline, ask for twenty genuinely different ones — different audiences, metaphors, angles. Models are tireless divergent thinkers; you are the curator. The final answer is usually a combination the model would never have ranked first.

   This inverts the usual instinct ("one polished answer, please") and it works because generation is cheap and judgment is yours.

   Try it this week

   Take something you are stuck naming or framing. Ask for twenty options across at least four different angles, pick the two best fragments, and combine them yourself.

2. Stage 2 · E2 PROFICIENT

   Make it play a part

   Models simulate perspectives on request: the skeptical buyer reading your proposal, the confused new user hitting your docs, the regulator reviewing your claim. Before a high-stakes meeting, have the model attack your weakest arguments and rehearse your responses — cheaper than learning the objections live.

   The craft is specificity: "play a CFO who thinks this project is a cost center and has been burned by two failed tools" produces sharper opposition than "play devil's advocate."

   Try it this week

   Before your next important presentation, brief the model as your hardest audience member and let it raise five objections. Count how many you had no answer for — those are your prep list.

3. Stage 3 · E3 DISTINGUISHED

   Stress-test and transform

   Use AI to generate what you cannot easily imagine: edge cases for the plan, failure scenarios for the launch, the awkward questions the FAQ does not answer. Adversarial generation finds holes while they are still cheap to fix.

   And exploit transformation range: messy notes into a decision matrix, a policy into a checklist, a dataset into the five anomalies worth investigating. If the conventional approach is "reformat by hand," there is usually a transformation ask that does it better.

   Try it this week

   Take a plan you own and ask the model for the ten most plausible ways it fails, ranked by likelihood × damage. Fix or mitigate the top two. That is creative use with a receipt.

4. Stage 4 · E4 EXCEPTIONAL

   Invent the move, then judge it like work

   At the top of this skill, you compose the moves: a simulated stakeholder panel reviewing three strategy options; synthetic customer feedback to pilot a survey before fielding it; the model arguing both sides of a decision while you referee. The inventiveness serves an outcome — it is never novelty for its own sake.

   And you keep the examiner's discipline: creative outputs get judged against reality like any other work. A brilliant simulated objection still gets verified with a real stakeholder; synthetic data never quietly stands in for the real thing.

   Try it this week

   Design one unconventional AI use for a current problem — something beyond drafting or summarizing. Run it, then write three sentences: what you did, why the conventional approach fell short, and how you judged the result. (That structure is, verbatim, what the exam asks.)

Ready to see where you stand? The free check scores all six competencies in about fifteen minutes.