Client Introduction & Business Requirement
Our client is a leading Indian tobacco manufacturer. They wanted to create a conversational AI solution that can help analysts get insights on datasets available on their competitors, and get queries answered related to products and financials.
The resulting insights should also be rendered using appropriate visuals and charts. This is to avoid redoing data models and reports, every time a new insight is required.
Solution Overview
- The solution involved creating a custom application that had the capability to accept conversational inputs (speech) from a browser interface or through a device using digital assistants like Alexa/Cortana.
- The user can ask any questions using conversational language on the datasets and entities available in their data warehouse. The user intents are deciphered using Azure Cognitive Services.
- The solution parses these intents and entities and converts them into queries through microservices.
- It then renders appropriate visuals (charts/graphs) using embedded PowerBI.
Business Benefits
- Support for conversational queries.
- Smart visualization of data without having to predefine metrics or report layouts.
- Faster response time to insights by slicing and dicing various cuts of data, based on evolving business scenarios.
- Highly responsive to executives and significantly cuts down time for decision making (versus) the traditional method of asking the questions and waiting for a project to get the reports built out.
