Prompt engineering is a growing area of focus in the field of AI. As more and more companies and individuals seek to enhance the output of their AI tools, prompt engineering has become a key area of interest. In this blog post, I will explain prompt engineering is, how it can be incorporated into your work, and provide examples of its use.

Prompt engineering is the process of designing the input text for an AI system in a way that helps it generate the desired output. It’s like giving the AI system a set of instructions or a question to guide it in generating a specific response. By carefully crafting these prompts, we can improve the quality and relevance of the AI system’s responses. So to say it even simpler terms:

Prompt engineering is like giving an AI system a set of instructions to help it give the right answer.

With your extensive knowledge of your job’s domain, you’re undoubtedly the most qualified person to create effective prompts for AI tools. As for me, I’ve developed a collection of predefined prompts that I use as templates whenever I need to utilize AI tools.

Having spent the past few months experimenting with various AI tools and poring over countless blog posts, I’ve discovered that several key components contribute to crafting an effective prompt. By providing specific details and clear instructions, a good prompt allows AI to produce accurate and relevant results.

Not any revolutionary stuff, but I find that including the following elements increases the results dramatically.

• Persona
• Context/problem
• Steps
• Constrains
• Output

One example could be this be that I want to know more about how a given topic in IT. Let us use DevOps as an example.

Persona: Act as a experienced IT person, having worked in various roles such as software developer, project manager and infrastructure engineer. Over the last years you have focused only on DevOps and have become highly skilled DevOps engineer. You have extensive knowledge about the principals behind DevOps.

Context/problem: I want to know about the different principals behind DevOps and get a beginners level of understanding.

Steps: Create a brief summary of the principals using the Parato principals . Condensing the key elements into sections based on the different principals.

Constraints: Don’t rewrite the principals or change them in any way. Use the book Accelerate by Nicole Forsgren as the main source.

Output: A summary of each principal explains to a beginner where every domain specific word is also explained.

Tips and tricks

Use the Parato Principle: Identify the 20% of [given topic] that will yield 80% of the desired results and provide a plan for gaining that knowledge.

Use the Feynman Technique: Explain [given topic] in the simplest terms possible as if teaching a beginner. Identify gaps in my knowledge of the [given topic] and create a study plan for them.

Make the tooling test you: Ask me some questions that are really challenging or give me some complex problems for the related [given topic] to test my understanding.

Storytelling: Transform the output into a summary containing the key elements and create an engaging story to increase the understanding and to help me remember it.

Summary

The AI in is simple form is stupid. So you need to help it help you in a kind of way. This is done by feeding it information, as much as you can and be very specific when doing so. Then you will achieve the best results.

As a skeptical tech person, I understand the concerns about the use of AI in everyday life and at work. While it is true that AI can be convincing in its output, it is important to maintain a critical mindset and verify the accuracy of the information it produces. It is always wise to research the facts and look for corroborating evidence before accepting any output as truth. By doing so, you can ensure that you are making informed decisions and using technology to its full potential while avoiding potential pitfalls. But the most important thing is getting started using it.