What is this AI and why should you care?
Imagine AI as a brilliant assistant who never sleeps, is capable of analyzing data (lots of data!), generating ideas based on what it’s learned, and emerging as a tool that can automate complex tasks.
At the end of this article, I’ll give you three practical prompting tips you can use today to get better results from AI.
For business leaders, generative AI is like a super-smart conversationalist who knows a lot of things. Large Language Models (LLM’s) that power generative AI are like the chef in a magical kitchen , they create responses by combining small pieces of information (called tokens) based on patterns they’ve learned.
Why care? Simply put, generative AI can take a lot of the content work that you need to do and streamline parts of it. What used to take you, and your colleagues, days of work can now be done in minutes with AI support. This is a huge productivity boost and capacity release that people are craving.
For example, it can draft marketing copy, analyze customer feedback, generate training materials, and in some more advanced cases, it can assist with automating repetitive tasks, saving time and money.
There is a catch - like any technology, it’s not a magic, silver bullet and there is no free lunch.
To get the best results, you need to give clear instructions, called “prompts”. Think of prompting as telling the chef exactly what dish you want. A vague order gets you a generic meal; a specific one gets you what you were looking for.
Here are several prompting tips you can use today to get better results from AI:
Imagine the LLM as “smart about the world, but not about your specifics.”
- They know general concepts but not your unique business context.
- Provide enough background information in your prompts so that the LLM can reason about the words you use.
- I find copy-pasting relevant documents, policies, or guidelines into the prompt helps a lot with LLM response quality.
Start the prompt with what the task is going to be.
- How many times has a colleague started telling you information and you had no idea where they were going with it, what they wanted you to do (if anything?), and what the end goal was?
- Be clear about what you want the AI to do with the information you provide later the in prompt.
End the prompt with the context dump, such as relevant documents or policies from point 1.
- This sets up the most recent attention of the LLM and gets them focusing in the right direction.
Example Prompt Structure
Your task is to create a compelling product description for our new tool based on the following company guidelines and product features. Ensure the description highlights the unique selling points we have which are X, Y, and Z. It is important to appeals to our target audience, which are primarily A, B, and C.
Ensure the description is engaging, concise, and highlights the key benefits of the product.
Here are the company guidelines:
[Insert company guidelines here, simply paste them as-is]
Here are the product features:
[Insert product features here, simply paste them as-is or upload a PDF if supported]
In the above prompt structure, you can see how we start with the task, providing clarity of what is coming in the next prompt. This gives the LLM transformer process the best chance of focusing the attention on the things that matter in the myriad of words you will provide later in the prompt.
Then there’s the guidelines about what you want and the prompt wraps up with the context.
This is a great, general purpose prompt structure you can use for many different tasks. It won’t be perfect for every use case, but it’s a solid starting point that you can iterate on.
The tag #ai-for-busy-people is a series of articles designed to guide business executives through a learning journey about AI, Large Language Models (LLMs), and prompting.
My aim here is to empower you to understand AI and apply it effectively in your businesses.