Role, Objective, Scenario, Expected Solution, Steps — A structured, solution-oriented framework for consulting-style problems.
ROSES is suited to problems that need a considered recommendation rather than a straightforward answer — strategic decisions, tradeoff analysis, or anything where you want the AI to reason like a consultant working through a specific scenario toward a defined kind of solution.
The expertise or perspective the AI should reason from.
The underlying goal behind the request.
The specific situation, including relevant constraints and stakeholders.
The type of answer you're looking for — a recommendation, a comparison, a decision.
How you want the reasoning or the answer broken down.
Bad prompt:
“Should I raise my prices?”
Structured with ROSES:
Role: You are a pricing strategist for small subscription businesses. Objective: I want to increase revenue without losing too many existing customers. Scenario: I run a $15/month tool with 400 subscribers, churn is currently 5% monthly, no price increase in 2 years. Expected Solution: A specific recommendation, not a list of general pricing strategies. Steps: State the recommendation first, then the reasoning, then one risk to watch for.
A specific price-increase recommendation stated up front, followed by the reasoning tied to the given churn rate and subscriber count, and a named risk to monitor — not a generic list of pricing tactics.
ROSES works well with Claude for nuanced, tradeoff-heavy business reasoning, and with ChatGPT when you want a faster, more direct recommendation.