saloniraj / prishu

Prishu is a verbal assistant that provides a hassle-free, one-to-one interaction with their users to provide a personalized response to the Search a product/ an easy way for Category navigation.

Due to the crises of Coronavirus Pandemic 2020, a large population prefers to stay at home in order to avoid/ minimize the risk of getting infected; which promotes the idea of online shopping using an e-platform to shop items of their needs. This project focuses to create a Conversational Solution which will enable the platform and its customers to discover and order products on vendors virtual platforms easily using customer's voice as the Natural Language Input, which also help in people of all age (like Young People in a hurry or old people who might need help to interact with website) to be able to have an easy experience over the platform.

Overview:

Prishu is a verbal assistant which provides a hassle-free, one-to-one interaction with their users to provide a personalize response to the Search a product/ an easy way for Category navigation. This helps each customer to receive/ experiences relevant results and operations via. Verbal Input. While most personalization tools require large data collection pre-emptively. However, in this project we wanted to focus on making a solution in which the Conversational Dataset grows with time and users, and to provide vendor a solution which is easy to start and deploy and better dataset over a longer period of time.

Datasets Used:

This project requires two datasets:

1. Conversation Dataset: Contains the information serving as an increasing data set for the Conversational Bot.
2. Product/ Item Database: Containing the details for the Item/ Products and their SKU (Stock Keeping Unit) sold by the Vendor

Technologies/ Frameworks Used:

1. Python 3
2. Machine Learning:
3. Natural Language Processing:

Use case and High Level Working:

The project initializes with a "Welcome Note" to greet the user after which it waits for the user input. After receiving the query, the bot processes the request and decides whether the request is a part of a conversation, or a product search request, or a command which Bot can perform. The program then chooses the closest matching response and communicates it back by printing the response (for logging and as a reference for the user for their previous requests) and telling it to the user as well.

Future Integrations:

This project can be further incorporated to be able to work with virtual assistant to make the platform much intuitive and easy for the customer to empower them to voice out their query to their voice assistant and can process query for them, without the need for customer to even open the Website/ App and increase the ease and reach of the Service.