Combine multiple techniques into a single powerful prompt. Layer role, audience, examples, and format together.
Each technique improves your prompt in a different way. But you don't have to choose just one — stack them together. A single prompt can set a role, specify an audience, show examples, and define a format all at once.
Think of it as: Role + Audience + Examples + Format + Constraints = One precise prompt
Here's how individual techniques combine into a single prompt:
Each technique solves a different problem. Role sets expertise. Audience sets depth. Examples set style. Format sets structure. When you stack them, you're giving AI all the guidance it needs in one shot.
A stacked prompt narrows the space of possible outputs. Instead of getting "any explanation of async/await," you get one that matches your specific needs across multiple dimensions.
Think of it like ordering food: "A sandwich" gives you anything. "A turkey sandwich on sourdough, toasted, with mustard, for here" gives you exactly what you want.
Pick the techniques that matter for your task. Write each one as a sentence or paragraph. Combine them into a single prompt. The order usually doesn't matter — clarity does.
Some combinations work particularly well together:
Perfect for tutorials and documentation. Role brings expertise, audience sets the level, examples set the voice.
Great for analysis and reviews. Role sets the lens, format ensures scannable output, step-by-step ensures thoroughness.
Ideal for batch content creation. Examples establish the pattern, format ensures consistency, constraints keep scope tight.
Best for advice and recommendations. Role brings expertise, audience personalizes it, constraints focus the scope.
You don't need to use every technique in every prompt. Pick the ones that address your biggest concerns. Build up as needed.
While order doesn't strictly matter, leading with the role or the most critical constraint helps set the frame for everything that follows.
After all your technique layers, end with the actual task. All that context should flow into the final instruction: "Now, write the tutorial."
Stacking combines techniques in one prompt. You get one response that incorporates everything.
Chaining uses multiple prompts in sequence, where each prompt's output feeds into the next.
Use stacking when you want a single, well-shaped output. Use chaining when you need to review intermediate results.