How to get the most out of genAI tools

Are good prompts enough to use genAI tools effectively, or is there more to it? As with any new technology, it is important to understand both the possibilities and the limitations. In this article, we will take you into the fascinating world of genAI and show you how to use this digital assistant as a powerful tool in your work.

Why you shouldn’t overestimate genAI tools

While tools like ChatGPT and Claude can produce very impressive results, we shouldn’t overestimate them either. Generative AI applications process your input and then produce a response. But they have no real knowledge of our world. What they do do is constantly predict the most logical next word. They do this thanks to complex neural networks in LLMs, or Large Language Models. So they are masters at pretending to understand our world as we humans do.

They also struggle with limited memory. In a new chat, the tools do not take into account what you have asked before. This can hinder the continuity of complex conversations. Also, remember that you remain the expert in your field. Think of the genAI tool as an assistant or even a student who can’t replace your in-depth knowledge and experience, but who can help you with executive or repetitive tasks. So keep a critical mind!

Which genAI tool is the best?

The answer may be disappointing: no AI tool currently offers a complete solution. The ‘best’ tool is the one that comes closest to your desired result. Here are the most commonly used tools and their pros and cons:

  • ChatGPT is the most popular in the list, working quickly and being suitable for many applications. It can also sound a little too generic at times.
  • The (free) version of Copilot that comes with your Office package currently offers the most security, but it changes its interface and options very frequently.
  • Gemini integrates seamlessly into your Google environment and has great access to data, but is not yet widely available and has limited multimodal capabilities (e.g. visuals).
  • In turn, Claude is very good and creative with complex information. This tool also does not excel in multimodal capabilities.

The basics of a good prompt

Once you have chosen your preferred tool, it’s time to write your prompt. If you keep a few best practices in mind, the output you generate will be better. Here are 3 basic tips:

  • Define the role, purpose, desired outcome, and format of the desired output and information.
    For example: You are an expert in… I want (not)… Provide me with… In this format… Use this information and examples…
  • Strike a good balance between begin brief and adding sufficient detail in your prompt. A longer prompt is not always better.
  • Don’t be satisfied with the first result you get from the tool. For example, ask it to evaluate the output or analyze what could be done better.

Conclusion

By applying the knowledge from this article and by constantly experimenting, you can harness the full power of AI tools. But remember that these tools, no matter how impressive, are designed to assist, not replace. Your unique insights, creativity, and expertise are still indispensable. As Claude himself so eloquently puts it:

“GenAI tools are like advanced brushes in the hands of an artist – it’s up to you to use them to create your masterpiece. Stay curious, keep learning, and discover how AI can take your work to the next level.”

Want to know more about prompting or genAi tools like Copilot?

Share this story

Let's talk about your next project.

Team Xylos is ready to meet you!

Other interesting stories

Learning Solutions

7 techniques to improve your prompts

Modern Workplace

How to handle large data sets (+2K) with Power Apps

Modern Workplace

Common Microsoft 365 governance challenges and how to overcome them

Corporate

5 prompts for Copilot that you didn't know you needed

Learning Solutions

Copilot through the eyes of our Digital Coach

Modern Workplace

Power Apps: parsing data from Power Automate