Topic 4: 7-Step approach to create a draft roadmap
Getting better results from Generative AI isn’t about finding a single “perfect prompt.” It is about working with AI more deliberately: giving it a clear task, providing useful context, setting boundaries, specifying what you need, and then reviewing and improving what it produces.
Think of prompting as an iterative working process, rather than a one-time instruction.
A 7-Step Approach for Better Results
1. Define the Objective: Start with the outcome you need. Be specific about what you are trying to accomplish, not just the topic.
2. Provide Relevant Context: Give the AI the information it needs to understand your situation. This could include background, documents, data, audience, business context, previous decisions or examples.
3. Set Constraints: Tell the AI what boundaries matter. Specify what it should include or exclude, the scope, assumptions it can make, tone, rules or anything it must not do.
4. Define the Output: Explain what a useful result should look like. You might ask for a table, summary, recommendations, user stories, checklist, comparison or presentation outline.
5. Use Examples When Helpful: If you have an example of the structure, terminology or standard you want, share it. Examples can make an ambiguous instruction much clearer.
6. Review & Challenge: Don’t accept an output simply because it sounds convincing. Check the facts, assumptions, evidence, gaps, bias and relevance. Ask questions such as: What assumptions did you make? What might be missing? What evidence supports this?
7. Refine Iteratively: Treat the first response as a starting point. Add missing context, correct assumptions, challenge the response and ask the AI to improve specific parts.
Apply It to Your Everyday Work
You can apply the same approach to many professional tasks. For example, a Product Manager might use it to create a draft product roadmap based on customer priorities and business goals. A Business Analyst might use it to turn requirements into draft user stories with acceptance criteria.
The quality of the result depends not only on the AI model, but also on the clarity, context and judgement you bring to the interaction.
Remember: It’s a Loop
Working effectively with GenAI rarely follows a straight line from prompt to finished answer:
Prompt → Review → Refine → Repeat
Each iteration gives you an opportunity to clarify what you need and improve the result.
The big idea: Don’t just prompt. Direct, contextualise, constrain, review and refine.
