using Azure Open AI (chatgpt) in power virtual agents to query SSAS tabular models
Budget: $250 – $750 USD
Agenda: To be able to provide chat bot experience for the users to query the SSAS tabular models in natural language and bring back the results from the tabular model.
Below is my thought but doesn't necessarily have to follow the same approach.
1. Pull the tabular metadata (Table, columns, list of measures) into file.
2. Train/Prompt Azure Open AI services with this metadata so when the users hit this model with question, it gives us appropriate query based on metadata provided instead of generic query.
3. Use Power Virtual Agent to provide chat experience to the users and call Azure Open AI service to pull the relevant dax query for the cube and then trigger power automate flow to pass the dax query to the cube and fetch the results back to user.
4. Publish the chat bot as web app.
SSAS Tabular Model Version - 1500 compatibility.
Below is my thought but doesn't necessarily have to follow the same approach.
1. Pull the tabular metadata (Table, columns, list of measures) into file.
2. Train/Prompt Azure Open AI services with this metadata so when the users hit this model with question, it gives us appropriate query based on metadata provided instead of generic query.
3. Use Power Virtual Agent to provide chat experience to the users and call Azure Open AI service to pull the relevant dax query for the cube and then trigger power automate flow to pass the dax query to the cube and fetch the results back to user.
4. Publish the chat bot as web app.
SSAS Tabular Model Version - 1500 compatibility.