Are you tired of sending prompts to the AI and receiving responses that make you wonder if it really understood you? Fear not, I'm going to teach you how to write prompts that are so clear and effective that even the most clueless model will get the idea.
PROMPT STRUCTURE
Each expert in prompting has his or her method, process or secret formula on how to structure the prompt in the best way. But all of them agree on a few points:
Give it a personality. Tell them who they are, what knowledge they have and how you want them to act.
Ex: you are an experienced digital strategist with a background in social media and more than 15 years of experience. You act as a community manager for a company that sells luxury shoes online.
Give it a task. Tell it what you need to get, the list of tasks to be performed.
Ex: analyze the posts and extract the brand's buyer persona following the "Jobs to be done" methodology, explain each point and include an example post.
Give an explanation. It usually works best if you describe the goal you want to achieve with the result.
Ex: you need this information to be able to create your new social media campaign, especially on Instagram, focused on a younger audience.
Give some guidance on how to respond. Include examples and indicate how you want to structure the response in terms of format and order.
Ex: present the information in orderly fashion, make a table to display the information and end the response with a list of recommendations for my campaign.
Example of using the prompt in a content view with Netflix's 202a Instagram, Youtube and Facebook posts.
GENERAL TIPS FOR TALKING TO GENERATIVE AI
🎯 No beating around the bush!
AI models don't need courtesies. Forget about "please" and "thank you".
Example: "I would like you to analyze the engagement rate of the posts".
Improved example: "Analyze the engagement rate of the last 50 posts on Instagram and summarize in four brief points the most relevant conclusions".
📋 Break down the task
Break down and make a list of the tasks you want it to perform, in chronological order if necessary.
Example: "Optimize social media strategy".
Improved example: "Identify the three best types of content based on engagement rate, extract insights and, finally, suggest ways to improve the content with those insights."
👍 Use positive language
It is always best to ask for what you want rather than what you don't want. Formulate requests in an affirmative manner.
Example: "Do not use technical jargon in the analysis".
Improved example: "Use clear and simple language to describe the analysis".
💁♀️ The power of examples
Use examples to make sure your answer will be optimal.
Example: "Make it informal".
Improved example: "Use an informal tone with words like "great," "come on," or "take a look."
📝 Organize your request
Organize the tasks you want it to perform and define them in as much detail as possible but without beating around the bush. You can use capital letters to reinforce some concepts.
Example: "Analyze this data".
Improved example: "Quantitatively analyze the post data and present the results in a table with three columns: brand, number of posts and total interactions."
😌 First and foremost, be clear
Remember, you're talking to a machine, so use clear and specific language. You need to say exactly what you want.
Example: "I want help with my project."
Improved example: "Describe three specific strategies to increase the engagement rate of the publications analyzed.
And finally...
Remember that every time you write a prompt, you are programming a mini adventure. Be patient with the results and edit the prompt as much as you need to until it gives you what you want.
There you have it! With these tips, you're ready to create prompts that get exactly the answers you're looking for, without misunderstandings or frustration. Unlock your new skill and discover a whole new world... 🤓
START PRACTICING 🤓
Ready to get started? Here are 9 examples of different structures to create your own prompts. We've made it as simple as possible for you. All you have to do is get creative! What are you waiting for?
