When you work with AI, formatting your prompts properly is what makes the model give accurate and useful answers. Below are the different prompt formats that will help you use AI effectively in the everyday tasks of your business and your job.
Question and answer
This format is simple: you ask a question and the AI answers it. It works well for getting direct, accurate answers to specific questions on any topic.
Example:
How do I increase conversion on a website?
Prompting with several examples
In this method you give the model several examples so it understands the task better. The method is useful for tasks where context or precision matters.
Example:
Answer this question:
I did not like that you charge twice as much for delivery as your competitors while offering no particularly interesting features. I will not be using your service any more.
**
How should you answer?
Example 1:
Customer review: I liked your product, but delivery was very slow.
Reply: Thank you for your feedback! We are glad you liked our product. We apologize for the delay and are working on improving our delivery speed. We hope your next experience will be a better one.
Example 2:
Customer review: The product does not match the description. I am disappointed.
Reply: Thank you for your feedback. We are very sorry the product did not live up to your expectations. Please contact our support team so we can help you resolve this.
Example 3:
**Customer review: Everything went great, I will order again!
Reply: Thank you very much for your positive review! We are glad you are happy with your purchase and we look forward to your next order.
Multiple steps
Here you give the model a task made up of several steps, and the model works through them in sequence. This is useful for tasks that require a chain of actions.
Example:
Task: Help me build a marketing strategy for launching a new product.
Step 1: Define the target audience.
Step 2: Develop an advertising campaign plan.
Step 3: Set up a content marketing strategy.
Step 4: Prepare a performance analysis plan.
Prompting with context
This prompt format means that extra information is supplied before the main question to help the model understand the task better. The method is useful for problems where background or a specific situation matters — analyzing customer data, for instance, or giving recommendations based on context.
Example:
Context: My online clothing store has been running for 2 years. We recently launched a new winter collection and we need to increase sales. Right now our audience is people aged 25 to 40 who are active on social media and shop online.**
Question**: What marketing strategies can we use to increase sales of the winter collection?