Lesson 1 · Prompt structure
Why most prompts give mediocre results
The problem is rarely the AI itself — it is how the task is phrased.
A typical weak prompt
“Write a marketing strategy for my business.”
The model doesn’t know who you are, what resources you have, your market or your goal. You’ll get generic textbook advice.
Structure of a strong prompt
Context → Goal → Constraints → Process → Output → Check
- 1Context
- Who you are, business model, product, audience and what has been done already.
- 2Goal
- A concrete outcome: not “a strategy” but “a plan to win 50 customers”.
- 3Constraints
- Budget, timeline, team resources, what is off-limits.
- 4Process
- The step-by-step procedure the AI should follow.
- 5Output
- The shape of the answer: table, checklist, procedure or bullet points.
- 6Check· Self-check
- Ask the AI to check its logic and assumptions before answering.
Context: [Describe your business, model, product and target audience] Goal: [A clear expected outcome, e.g. a plan to test 3 B2B lead acquisition channels] Constraints: [Budget up to $1,000, 1 marketer on the team, 30 days, no paid ads] Process: 1. Identify our audience's 3 main pain points. 2. Write a value proposition for each pain point. 3. Suggest 3 channels with the fastest time-to-value. Output: A table with columns: Channel | Hypothesis | Resources needed | Expected result. Before finalizing, check the answer for missing assumptions, weak logic and unclear recommendations.
Without structure
Ten pages of generic advice (“be active on social media”, “create quality content”) you can’t act on.
With structure
A focused action plan that fits your team’s resources, where each step is a hypothesis with a success criterion.