Leveraging Gen AI for generation of content for knowledge base is easy; the key is to get the instruction right.
when using the out-of-the-box Generative Answer feature in Copilot Studio, the responses tended to summarize and provide context rather than delivering the specific instructions users needed.
After a few trials and errors, this is a framework I generally try to leverage while crafting instruction for Copilot Studio chatbots. The prompt is structured to ensure clarity, accuracy, and proper formatting, making it easier to generate precise and useful content for a knowledge base
A well-crafted instructions ensure that the AI understands the user's intent and delivers the desired output.
1) Task: Clearly list all the steps required to complete the instruction. If you do not then the generative AI will basically summarise what it thinks is the bast answer. Be articulate on how you would interpret the SoP. What are the segments that you would be interested to see from the documents.
2) Clarity: Provide clarity on the expected output, ensuring the AI understands what the final result should look like.
3) Refine: Added couple of instruction like
Validation:Ensure the AI validates the user query against the knowledge base accurately.
Formatting: Pay attention to the formatting of the markdown table to ensure it displays correctly.
TRC framework
I am sure we could optimise the number of steps in the prompt for the specific use case . Might need a lot of trial and error.
Example :
