Welcome back! In the first lesson of the course, you learned how formatting your prompts — using headings, lists, and code blocks — helps language models (LLMs) better understand your requests. Let's build on that foundation by focusing on another essential aspect of prompt engineering: defining constraints and requirements.

When you give an LLM a task, it will try to fulfill your request. However, if your instructions are too broad or vague, the model might produce answers that are off-topic, too lengthy, or not useful. By adding explicit constraints and requirements, you guide the model to deliver precisely the desired output.

In this lesson, you'll learn what constraints and requirements are, how to write them effectively, and how they can make your prompts much more powerful.

A constraint is a rule or limit you set for the LLM's response. A requirement is something the response must include or follow. Both help you control the output.

Let's look at a simple example. Imagine you want to create a survey for your customers. Here's a basic prompt:

This prompt is clear, but it leaves a lot of room for interpretation. The LLM might suggest questions about anything — pricing, product features, or even personal information.

Now, let's add some constraints:

Adding these constraints tells the LLM exactly what you want and what to avoid. This makes the output more focused and valuable.

• Organize constraints as a bulleted list: This makes them easy for both you and the model to read and follow.
• Be specific: Vague constraints are easy to overlook. Spell out exactly what you want or don't want.
• Limit the number of constraints: Too many constraints can overwhelm the model, causing it to ignore some. Focus on the most important ones.
• Be strict and explicit: Use strong language like "must" to emphasize what is required.

Let's see how these tips work in practice.

Suppose you want the LLM to write a product description but want it to be short and not mention the price. Here's how you might write your prompt:

Explanation:

• The first constraint sets a word limit, ensuring the answer is concise.
• The second constraint tells the model to avoid discussing price.
• The third constraint directs the content to focus on a specific topic.

By being clear and direct, you help the LLM give you precisely what you need.

Let's say you want the LLM to generate a list of creative team-building activities for a remote team, but you have some specific needs. Here's a prompt with well-defined constraints:

Note:

• This prompt also uses formatting to separate context, ask, and constraints.

In this lesson, you learned how to use constraints and requirements to make your prompts more effective. By being clear, specific, and direct, you help LLMs provide better, more useful answers.

You saw how adding constraints step by step changes the output, and you learned how to write your own constraints for different tasks. Next, you'll get to try these skills in hands-on practice exercises. This will help you get comfortable crafting prompts that get desired results. Good luck!