You might already know that prompt engineering is the practice of crafting effective commands or instructions (or “prompts”) to guide AI models like ChatGPT in generating desired outputs. Why is it important?
To explain this metaphorically: Have you ever wondered why, when the same tool is in the hands of different people, some create greater results in less time with it, while others struggle? What makes the difference?
It’s about how the tool is used and whether the right techniques have been applied. The same can be said about using ChatGPT. Proficient ChatGPT users who’ve mastered the art of prompt engineering are more likely to achieve their desired outcomes more efficiently and effectively, and make the most out of the tool’s capabilities.
In this post, we’ll explore 10 essential prompt engineering techniques that will set you apart from “ordinary ChatGPT users” when you apply them.
Table Of Content
1. Role-Assumption Prompting
The first technique is Role-Assumption Prompting, where you instruct ChatGPT to take on a specific role. This helps guide the AI’s responses to be more relevant and tailored to the situation.

When to Apply: Use this technique when you need the AI to generate responses from a specific perspective or expertise, such as a teacher, doctor, or marketer.
Example: If you want ChatGPT to handle a customer complaint, you could prompt it with:
Act as a customer service representative with more than 10 years of experience in the retail industry. How would you handle a complaint about a delayed order?
2. Chain of Thought
Chain of Thought is about prompting the AI to break down and show you its reasoning process step-by-step – instead of directly shooting you a response. Add the magic phrase “Let’s do it step-by-step” or other phrases to the same effect in your prompt to get ChatGPT to explain how it arrives at its answer.
When to Apply: This prompt engineering technique is particularly useful for complex tasks or when you suspect ChatGPT may be providing incorrect answers – or straight-up hallucinating. Use the Chain of Thought technique for detailed explanations and have it walk you through its reasoning, so that you can verify its responses and provide appropriate feedback.
Example: ChatGPT will show you the process of arriving at the answer when you prompt:
Explain why 2 + 2 equals 4. Let’s go step-by-step.
3. Zero, One, Multiple Shots
This technique involves guiding the AI with examples. “Zero-shot” means you provide no examples, “one-shot” means one example, and “multiple shots” means several examples. How do you decide how many examples to provide? You have to weigh the effort required against how specific you expect ChatGPT’s response to be.
When to Use: Use zero-shot for general tasks or when a general response is good enough, one-shot when the task needs a little more guidance, and multiple shots when you need a very specific response.

Example:
When you request ChatGPT to plan a meal for your cat, you can prompt:
- Zero-shot: “Create a list of healthy meal options for cats.”
- One-shot: “Create a list of healthy meal options for cats. Please refer to this menu.”
- Multiple shots: “Create a list of healthy meal options for cats based on these 5 menus my vet gave me.”
Following the prompt, you can send the examples/menus for ChatGPT’s reference.
4. Prompt Priming
“Set the stage” before you request ChatGPT to start working on a task by supplying background information or initial context. This is what Prompt Priming is about—you “prime” the AI (in other words, prepare it) to produce a response that is more aligned with your expectations.
When to Use: Use Prompt Priming when it will be helpful to set the stage for ChatGPT to carry out your task. This prompt engineering technique also helps the AI understand the context, so that it tailors its response accordingly.
Example:
I have a 5-year-old Siamese cat who’s highly intelligent and easily bored. He loves puzzle toys that challenge him to think and keep him engaged for long periods. Recommend a suitable toy for him.
5. Generated Knowledge
The Generated Knowledge technique is similar to Prompt Priming, in that they both set the stage before ChatGPT works on your task. The difference is, instead of supplying ChatGPT with the background information, you ask it to generate or recall relevant information or knowledge.
When to Use: This technique is helpful when you don’t have the required information to prime ChatGPT, or if you just want to verify the AI has a solid foundation before starting your task.
Example: Let’s say you want to ensure that ChatGPT is on the same page with you on a certain topic. You might first prompt:
Before answering, summarize the key principles of quantum physics.


After verifying that its knowledge is correct, ask it to incorporate this knowledge into its output. For example:
Based on your knowledge of quantum physics above, explain how these principles apply to quantum computing.
6. Prompt Chaining
Prompt chaining is the practice of guiding ChatGPT to complete your task over a series of sequential prompts, with each prompt building on ChatGPT’s response from the previous prompt. This technique might require thoughtful design in your prompts, and is ideal for complex tasks that are best approached in several steps.
When to Use: This prompt engineering technique works best for tasks that need to be broken down into multiple stages. Ensure that each step is thorough before moving on to the next.
Example: A blogger is getting ChatGPT’s help to create an infographic based on an article. She could start with:
First, summarize the article on climate change.
and then follow up with:
Now, use that summary to create a list of the top 5 key points.
She would continue this process step-by-step, with each step building on the previous one.
7. Ask Before Answer
Encourage the AI to ask clarifying questions before it attempts the main task with the Ask Before Answer technique.
When to Use: When you are not sure if ChatGPT has fully grasped your requirements, or when you suspect your prompts might not be clear enough or there might be missing information, ask ChatGPT to tell you what it understands about your task.

Example: Apply the Ask Before Answer technique by simply asking ChatGPT:
Do you understand the requirements? Is any other information needed?
8. Iterative Refinement
Let’s face it: ChatGPT rarely gets it right on the first attempt, especially with tasks that even humans can’t do in one shot. That’s where Iterative Refinement comes in. Apply this technique by putting ChatGPT’s output through a series of adjustments to gradually refine it until you’re satisfied (or, well, decide to give up).
When to Use: This prompt engineering technique is effective for creative tasks where you aim for a polished final product.
Example: Start with:
Write a draft of a short story.
Then refine it by saying:
Now, add more detail to the characters’ backgrounds.
followed by:
Polish the dialogue to make it more realistic.
Each iteration improves the story until it’s fully developed.
9. Prompt Templates
Over the years, experts in prompt engineering have developed their “winning formula” for prompting, known as “prompt templates.” These templates are typically one or two sentences with placeholders that regular users can reuse with slight modifications. See examples below.
When to Use: Prompt templates simplify the process of creating effective prompts by providing a ready-made and proven-to-work structure. Scour through existing templates to find ones that suit your prompting needs.
Example:
- Explain [Topic] to [Audience]
- List the [Number] most important [Category] for [Task/Topic]
- Compare and contrast [Item 1] and [Item 2] in terms of [Criteria]
- What are the potential challenges of [Action/Decision] and how can they be addressed?
- Summarize [Content] in [Number] sentences
10. Contrastive Prompting
Ask ChatGPT to compare and contrast different ideas, options, or scenarios instead of spending hours doing it yourself (it’s great at it!). That’s what Contrastive Prompting is about.
When to Use: Use this technique to explore different perspectives. Pro tip: Don’t just say “Compare X and Y” if you want detailed comparisons. Ask ChatGPT to list the pros and cons.
Example: Get ChatGPT to provide insights into the pros and cons to help you weigh your options with a prompt like:
Compare the benefits of remote work versus office work. List the pros and cons of each option.

Conclusion
In this post, we’ve covered ten essential prompt engineering techniques. By mastering these techniques, you can make your interactions with ChatGPT more efficient, effective, and tailored to your needs.
Experiment with these methods to see which ones work best for you. This will take your prompt engineering skills to the next level.
One Comment