How to create an Azure chat bot where the data source is multiple PDF files of similar data and the chat bot answers the question based on the filename provided in the query?

Question

I developed an Azure chatbot that searches through the user manuals of two similar devices,for example, the user manuals of two different TV models, the LG LM56 43 (108.22 cm) Smart FHD TV and the LG LM56 43 (108.22cm) FHD TV. Is there a way in which the user can mention the name of one device and the chat bot will answer the question based on the user manual of that device, even though there is the same topic in the user manual of the other device?

Answer 1

Hello @Kannan P S Yes, it is possible to design your Azure chatbot to handle your scenario.

One way to do this is by implementing a context-aware system in your chatbot. Here’s a high-level approach:

1. When a user asks a question, the chatbot should first identify the device being referred to in the question.This quickstart guide will help you understand how to implement Named Entity Recognition (NER) in Azure. NER can be used to identify the device names in the user’s questions.
2. Once the device is identified, the chatbot should set this device as the context for the current conversation. All subsequent questions will be answered based on the user manual of this device. This can be done with Conversational language understanding (CLU) which enables users to build custom natural language understanding models to predict the overall intention of an incoming utterance and extract important information from it. CLU only provides the intelligence to understand the input text for the client application and doesn't perform any actions.
3. If the user mentions another device in their question, the chatbot should update the context to the newly mentioned device. When answering a question, the chatbot should search through the user manual of the device currently set in the context.

Hope that helps. Let us know if you have more questions.

-Grace