Development of a Conversational Chatbot Indexed to a Database

Chatbot created using the Deepset platform. The chatbot is capable of maintaining a conversation while also querying a Snowflake database it is connected to. In this way, the client can query the database without needing to use a specific syntax, simply through natural language.

The queries made by the user are translated into SQL queries thanks to a generative LLM that resides at the core of the bot. This allows the query to be executed on the database and the information retrieved to answer the user’s request.