Practical AI: Prompt Build
UG Expert Kylie Kiser walks through more AI tools available to users, specifically how to build AI and write prompts.
Key Takeaways
- Prompts defined: Prompts are instructions given to the AI model. Users can define the structure and the output. The more clear the prompt is, the more consistent the results will be. The answers can be different each time which is okay in some scenarios but in others, there could be a lot of risk involved. So, it’s important to make the results as consistent as possible by including as much information and context as you can.
- Considerations: What am I doing? What is the task that needs to be done? What’s the format of the information be inputted and what format does the output need to be? Are there any constraints? These examples of questions you can consider when crafting your prompts.
- Starting simple: You don’t want to include real life examples containing any secured information. You can start simple with a basic taks and add structure as you go, testing and refining with each output.
- Building prompts: Prompts can be used within Power Platform, Copilot, Power Apps, Power Automate, and Copilot Studio. Kylie demonstrates what this looks like and how to build your own prompts. Users can take advantage of templates and different available models.
