D-R-E-A-M
An iterative 5-step prompting framework to get precise AI responses on complex topics, through a structured cycle of definition, refining and adjustment.
Description
D-R-E-A-M is an iterative prompting method in five steps: Define (set the intent), Refine (add constraints), Execute (submit the prompt), Assess (evaluate the response), and Modify (adjust and resubmit). This cycle repeats until you get a usable result. It is particularly useful for complex or strategic tasks where a single prompt is not enough: product briefs, competitive analysis, argument structuring. The Assess phase is often the most neglected, yet it is the most decisive.
Objectives
- Identify problems
- Foster innovation
- Structure AI prompts
How to apply D-R-E-A-M
- 1
D - Define the intent (Define)
Clarify what you want to achieve: what problem, for whom, in what context. Formulate your intent in one or two sentences.
- 2
R - Refine with constraints
Enrich your intent with constraints: tone, length, format, target audience, angles to avoid.
- 3
E - Execute the prompt (Execute)
Submit the prompt to the model. Let the response complete fully before intervening.
- 4
A - Assess the quality (Assess)
Evaluate the response: does it address the intent? What is missing or too generic? Note 2 to 4 areas for improvement.
- 5
M - Modify and resubmit (Modify)
Rephrase the prompt specifying what needs to change and what should stay. Resubmit and return to the Assess step until satisfied.