The Four-Step Exercise
Scenario planning was formalized by Pierre Wack at Shell in 1971 and used to outperform competitors through the 1973 oil crisis. The technique: develop plausible future states, test your plan against each. AI accelerates the slow part.
AI (a few minutes)
"Generate four plausible scenarios for how [situation] could evolve over the next two quarters: one optimistic, one pessimistic, one disrupted by a competitor, one disrupted by internal resourcing."
You (fifteen minutes)
Mark each scenario with what you actually know about current conditions. Add signal you have that supports or contradicts each future. This is your judgment, not the AI's.
AI (a few minutes)
"For each scenario, identify the one assumption in my current plan that would break first." Each scenario gets one critical assumption surfaced.
You (the rest)
Pick the most likely scenario based on what you know. Adjust the plan. This step is entirely yours. AI informed it; you own it.
Practical example
A product manager has 45 minutes to finalize a Q3 roadmap. She prompts AI for four scenarios for how customer adoption could evolve and gets them in two minutes. She spends fifteen minutes annotating each with what she knows, then prompts for the first assumption that would break in each. The AI returns four critical assumptions. She identifies the resourcing scenario as most likely and adjusts the roadmap accordingly. Total: roughly twenty minutes of AI interaction plus twenty-five minutes of her judgment. Without AI, she would have skipped the exercise entirely.
Apply It Today
On one real plan, run the four-scenario prompt and then ask which assumption breaks first in each. Adjust one thing based on what you find.
Reflect