One of the things that makes me happy and boosts my mood - believe it or not - is to find innovative ways to do things and alternatives to the norm that can leverage work on so many levels. While navigating this scenario I discovered a piece of technology I love: AI prompts within Power Platform. My friend Daniel Laskewitz delivered a session at Dynamics Minds about the topic and I knew I had to do something with this. He kindly shared some info, so if you want to know more go and have a look at his work.
You might be wondering what AI prompts are...
In the world of AI, prompts act as guides for generative models. These prompts are essentially a set of instructions that tell the model what kind of task to complete.
Shall we dive into the wonders of AI prompts within Power Platform?
Below you will find a summary of what AI prompts are within Power Apps, Power Automate, and Copilot Studio. These are ways to add AI to your apps, flows, or agents. Including AI - respecting the responsible guidelines - is one of the best things you can do with your existing or new solutions.
As in life, the clearer you are, the better. Building your prompt it's the same, be as detailed as possible, to get a good outcome you will need a clear, defined, organized prompt. Let's see an example of a "good" prompt and a "bad" prompt:
Do ✅ | Don't ❌ |
Based on the 'Approval Status' field in a SharePoint list item, generate a customized message notifying the submitter of the decision. | Use the prompt when a file is uploaded. |
Where can you find the AI prompts within Power Platform?
I recorded a short video where you can find the AI prompts within the AI hub, either in Power Apps or Power Automate, the path is the same one.
Below you can find six common uses where AI prompts can be used, have a look at them and the entities that can be added through your apps, flows, and agents.
Test before make
One of the cool things you can do with the prompts is to test before using them. Isn't that fun? Similar to going to a perfume store and trying out before buying, this is the same but less - or more - glamorous.
Summarize text
Extract information from text
Classify text
Sentiment analysis
Respond to a complaint
If you want to explore more, I highly recommend you check the following Microsoft resources:
Comentarios