Let’s talk about the importance of prompt design when using any AI tool, specifically ChatGPT.
First, the prompt is the action you’re asking AI or ChatGPT to accomplish—simply put, it’s the task you give it.
The design is what goes into that prompt. Maybe it’s the specific tone you want to speak in or the task you’re giving it. It’s the purpose behind the content.
Let’s look at some examples!
Different Prompt Designs
Ensuring your prompts are well thought out is crucial so you get the most from the AI tool.
You can see below that OpenAI did a great job when we asked it to summarize the linked blog post in two hundred words.
But when you only give AI that first prompt, it won’t produce content in a voice that sounds like your own writing. So, for our example, we then prompted AI to rewrite the summary in an upbeat, fun, and informative tone. It’s helpful to add a more personal touch if possible.
This shows the importance of why you would want to add a tone to your prompts.
The next step for prompt design is regarding the tasks you give AI. We asked OpenAI to write something else in the same style. See below.
OpenAI wrote about Canva and how users can generate high-quality visuals. This content had nothing to do with the previously linked blog post we asked it to summarize.
This is a basic example, but it shows how important it is to specifically ask AI to do what you want it to do. We didn’t want to talk about Canva here; we wanted AI to write something else based on the previous summary in that same tone.
We tried another prompt in the same writing style because we enjoyed the tone and style as it matched ours:
Once you nail down the tone with AI with prompt design, you’ll save time editing what you get back.
You can see that AI perfectly executes what you ask it to do when you have specific prompt design instructions. It’s important to be clear in the prompts you give AI, specifically in OpenAI and ChatGPT. Doing so will help you achieve the exact output you’re looking for when AI responds to your prompt.
Editor’s note: This article was originally a transcript reworked by Sophia Stone, Interact Intern