Good Prompt Structure

Even with great ideas and inputs, a poorly structured prompt can lead to inconsistent or weak results.

This article covers simple patterns for structuring prompts so models understand what you want every time.

Think in sections

Instead of one long paragraph, break your hidden prompt logic into clear sections, for example:

1. Role / persona – who the model should act as.
2. Goal – what you want it to achieve.
3. Inputs / context – what information it will receive (and from which placeholders).
4. Steps or process – how it should think or work through the task.
5. Examples - give the model some ideas what you expect as an output
6. Output format – what the final answer should look like.

This makes the prompt easier to read, maintain, and debug.

Start with role and goal

Lead with a short role and goal statement:

You are a senior copywriter specializing in B2B SaaS.
Your goal is to write a clear, friendly email based on the information provided.

This sets context immediately and helps the model choose the right tone and level of detail.

Clearly describe the inputs

Tell the model what information it will receive and how it's labeled:

Given the following key information:
- The target audience: {{ target_audience }}
- The main offer or message: {{ offer }}
- Any constraints or extra notes: {{ notes }}

This reinforces what each placeholder means and reduces ambiguity.

Give step‑by‑step instructions

Models respond well to ordered instructions:

Follow these steps:
1. Briefly restate the main offer in one sentence.
2. Write a short email that speaks directly to the target audience.
3. Use a friendly, professional tone.
4. End with a clear call to action.

Numbered steps encourage more structured thinking and output.

Provide some examples

It's helpful to show the model sample inputs and expected outputs. This sets a clear target for style, length, or content. You can add one or two short examples after your instructions:

Example:

Target audience: Marketing Managers  
Offer: Free trial of our analytics platform  
Notes: Emphasize quick setup

Output:
Subject: Unlock Your Free Analytics Trial

Hi there,

Experience powerful analytics with no setup headaches. Start your free trial today and see insights in minutes!

Best,  
[Your Company]

Well-chosen examples help the model better understand your expectations and improve consistency in its responses.

Specify the output format

Tell the model exactly how you want the answer to look:

Output:
- A subject line on the first line.
- A blank line.
- The email body.
Do not include any explanations or extra commentary.

Being explicit about formatting makes your prompt easier to plug into workflows and integrations.

Keep structure stable as you iterate

When you improve a prompt:

• Try to keep the overall structure (sections and steps) consistent.
• Make small, targeted changes to wording or examples.
• Re‑test with the same inputs to compare behavior.

A consistent structure makes it easier for you—and the model—to reason about what changed and why.

Over time, you'll find that a clear, repeatable structure is one of the biggest drivers of reliable prompt performance.