Best Practices
Here is a list of best practices you must follow while designing a semantic model to get better clarity in response:
Ensure that you provide proper tags and descriptions for a clear business context. Column level description should be leveraged to provide information like:
Sample 2-3 values for domain-specific attributes.
Defining special date time aspects like Holiday months, Weekdays/Weekends other than usual ones, Business season, etc.
Some special instructions you want Large Language Model (LLM) to follow, like not using specific columns for certain purposes. For example, order priority is used to identify returned orders.
Always have complete date attributes in the semantic model and a predefined hierarchy.
Hide all unnecessary Levels, Attributes, and Measures not required for conversational analytics.
Special characters should not be present in the semantic model’s name or tags and descriptions of the semantic model.
Ensure that you clearly define the Keys field for level and attributes for which you want a different sort order than the order of the display field.
The clearer the context of the question, the better the response will be. For instance, instead of asking, "Show sales for 143-435-564," you should ask, "Show sales for product ID 143-435-564."
Break down complex questions into smaller, more specific follow-up questions. For example, <>
Refer to the metadata of a semantic model to better frame your questions in alignment with the model's capabilities and the type of information it can provide.
Considerations
Kyvos copilot leverages Large Language Model (LLM) to convert natural language to SQL, which might sometimes produce inaccurate generation. Always verify the answer by checking the information in the answer. In case of an error or incorrect response, try regenerating the answer by re-submitting the question.
Copilot can't make layout changes. For example, if you tell Copilot to resize the visuals or to align all the visuals perfectly, it won't work.