On this page, you will find all the information on how we can integrate ChatGPT into MyAdventure Group.
MAG Team
Day-to-day operations.
Improve daily operations
ChatGPT is a chatbot based on large language models (LLG), that enables users to refine and steer a conversation towards a desired length, format, style, level of detail, and language.
Watch this video to have a thorough understanding of the Prime, Prompt, Polish model.
The PPP Method, introduced by Everyday AI, enables ChaGPT users to fully optimise their experience in creating prompts for specific topics with three key rules/methods to follow:
555 Rule
REFINE-Q
Polish
The 555 rule distinguishes when you should use ChatGPT and protect your ROTI. You can utilise ChatGPT for research, content writing, data analysis, etc.
Each chat's knowledge about a specific topic is too broad. It is critical to provide each chat with the necessary background information and examples to ensure high-quality prompts.
Things to consider:
Will it take you more than 5 minutes without ChatGPT?
Can you reuse the same prompt at least 5 times?
Can you verticalize the prompt in 5 different areas?
ROLE - Provide the role that you want GPT to play. Explain in detail the experience and background. You can also include insights from other professionals in the same field.
EXAMPLES - Provide examples of inputs and outputs for the role.
FETCH - Provide websites from where GPT can fetch information. You can include examples of similar competitor websites.
INSIGHTS - Ask GPT to provide/include specific information from certain websites.
NARRATE - Explain to GPT the specific key points you want it to include for the copy.
EXPLAIN - Explain what kind of copy you want GPT to come up with.
QUESTION - Do not ask for an output. Ask if it has questions related to the role, audience, competitive landscape, or any specifics related to the task.
After asking your question, ensure that you answer each question in the order it was asked.
Once you have responded to all the questions, write down at the bottom of the page:
"Based on my answers and your additional research, please ask me all of your remaining questions or any clarifying information you need from me before I provide you with the prompt."
NAVIGATE - State what's good and what needs to be improved.
STEER- Rewrite the prompt and give specific examples of good points and areas of improvement. Say “Here’s an example of a good prompt” or “Here’s how you can improve your content..” “Instead of saying..”
FINISH - Refine the copy one last time, and make sure you are happy with the tone of delivery and the content that’s been produced. And finally, end the chat.
After completing this process, you will have successfully trained an AI Bot to generate prompts that do not sound like they were made by a machine.
Expert Chats
1 chat for 1 specific skill set
Train for a VERY specific purpose
Train as many chats as possible and keep them SEPARATE
REFINE REFINE REFINE - utilise the REFINE-Q method.
Create a labelling system
Memory Recall
It is when you ask GPT to summarise all the information it has learned so far.
When you want to add new information or notice that it's missed something, run another recall or re-prime.
*Ask questions like:
"Please remind me of your background and expertise for this chat, and how you will write this copy"
Workflow Best Practices
Double-check which version of ChatGPT you’re using.
*The paid version (ChatGPT 4) offers a better experience when utilising the PPP Method.
You can have unlimited chats open for the same account
You can either split your screens or use multiple screens
Each chat can only be in one mode. You can’t switch. Choose from 3.5 (Free) or GPT 4.
If you see a warning that you’ve reached a limit for your chat, don’t switch to default mode. Don’t click that if you’ve trained your expert chat.
Last Update Date: 15/05/24
Updated by: Eyo Faithe