Prompt Engineering

When you use a tool like Microsoft Copilot, you are interacting with a Large Language Model (LLM). A prompt is the input text that you use to instruct the LLM what task to perform, guiding it to produce a relevant output. Crafting clear, well-structured prompts is needed to get good results from LLMs. This is where prompt engineering comes in – it’s the process of designing prompts that communicate your intent to the LLM and help it generate the kind of output you are looking for. 

What makes a good prompt?

Here are some key things to keep in mind when writing prompts: 

• Be specific: Tell the AI exactly what you want it to do. Use clear action words like “explain”, “compare”, “summarise”, etc. 
• Give context: Help the AI understand what you’re asking by providing background info or examples. This helps it give more relevant and useful responses. 
• Set guidelines: Let the AI know what you expect in terms of things like length, format, style, or perspective. This keeps it on track. 
• Iterate: Don’t expect a perfect response on the first try! Start with a basic prompt and then keep refining it based on the AI’s outputs until you get what you need. 

Here are some examples of prompts:

Summarise the key points of the following paragraph in 5 bullet points. Focus on the main argument and supporting evidence: [insert paragraph] 

You are a social media expert giving advice to a small business owner who wants to use Instagram to promote their products. Write a short, friendly 200-word post sharing your top 3 tips for creating engaging Instagram content. 

See how these prompts are clear, provide helpful context, and give the LLM a good idea of what kind of response is expected? That’s the key to prompt engineering!