Master the Art of ChatGPT Prompt Engineering for Stellar Product Management

Master the Art of ChatGPT Prompt Engineering for Stellar Product Management

The Basics of ChatGPT Prompt Engineering for Product Management

Follow me for more weekly AI content in relation to product, I write about AI adoption methods.

1. Think like a computer:

Just like any system, understanding the root of how the system works will assist in its use.? ChatGPT is a low level elementary AI that is dynamically searching for content that best matches your prompt.? The fun of using ChatGPT is remember that you’re talking with a computer that is looking at all other peoples responses to the question or like question you have asked to present an acceptable answer.? With that in mind, remember that not all information is going to be useful or correct.? This is the biggest Achilles heel of ChatGPT, you must have an idea of the outcome or understanding enough of what you’re asking to see if the answer makes sense.? The hilarious part about using ChatGPT is how transparent articles, emails or written documentation that is copy/pasted directly out of ChatGPT without any form of validation.?

The best method to get good responses is using action oriented verbs that define the intended request that identifies the origin and target.? It’s very important to make sure that all requests have an intended audience or target otherwise the response will have limited relevance for what you need.? Consider adding as many elements as possible to the prompt plus defining the intended length of the response.

The basic formula is:? Define the request with action words and an audience, define the audience, and what the intended document should look like or length.

As example, let’s try to make a prompt for creating a product discovery document.?

“ChatGPT, I want to create a product discovery document that will allow me to document product requirements with a non-technical audience.? The non-technical audience is a combination of sales, marketing and accounting people working with in the software engineering industry.? The document should include a section for: Minimal viable product, resources needed, definition of success, estimated budget (Capex and Opex expenses) and defining who in charge of each element of the product (who is in charge of what).? Also include other sections that could be relevant for best explaining a product request.? This document should be less than 1 page long and include a cover page.”

So how can we merge the human with the computer to help with our prompts?

2. Give as many details as possible or… ask ChatGPT for help:

The hidden feature of ChatGPT is that you really don’t need to know everything before starting a prompt.? Have an idea of what you would like to accomplish but…why not ask ChatGPT to help define your prompt?

How about we ask ChatGPT for help making the prompt to ask for the above product discovery document again.

“Hey ChatGPT, I want to make a document that will help me define product features but I need to ask questions to a non-technical audience who might not know fully what they want?? What are 5 questions I can ask to get the conversation started?”

ChatGPT response with 5 questions will follow.? Now we have to ask ChatGPT to merge my first request with the 5 questions to create a discovery document.

“Hey ChatGPT, can you please combine my first prompt with the 5 questions response to generate a product discovery document that will allow me to ask my non-technical audience to best define their product requirements.? The document should be about a page long and include a cover page.”

Now we have combined both the first prompt and the response to create the document we needed.? However, just because we did this once…doesn’t mean we can’t keep asking ChatGPT to clarify and iterate!

3. Each conversation builds on itself – The Joy of iterative generative AI prompting!

Anytime that a conversation with ChatGPT is asked a question, this creates a new conversation till continues to build on every response given within that conversation.? With the above document we started to ask for, its completely possible to dial in the document to fit your purposes by correcting the response with direct feedback.? What am I talking about?

“Change the section length from 5 points to 3 points”
“Rewrite the title to be more catchy for a sales oriented audience in the pharmaceutical industry”
“Reorder the sections in reverse order”
“Translate this document into Spanish”

Feel free to get crazy with asking more clarifying questions or prompts within your conversations.? Also, its possible to edit previous responses if you don’t like the response you got.? Click on any of your responses and there is an edit feature to corrected the response or keep going on the conversation.? Another method is to copy paste the elements of responses that you like and tell chatgpt to use this style of writing for the rest of the responses in the conversation.

These are a three interesting ways to not only use prompts but to describe what is needed from ChatGPT.? Always remember that like any skill, prompt engineering is a process that will develop for you as your use increases with ChatGPT and other AI prompting services.?

If you like this type of content, give me a follow here on LinkedIn and this article was written by a living breathing human ;)

Beau

要查看或添加评论,请登录

社区洞察

其他会员也浏览了