AI Powered Chat bot

Overview

A software development company developed UI for a chat box. Later they realised it was a good idea to train the chat bot as this wouldn’t need any resource answering at least the general queries. The business manager decided to hire a data scientist who could train the bot to chat with the resources.

Challenge

The data was limited as the software development company did not maintain it. A market research company was employed to collect the data related to travel industry. A data scraper was employed to scrape the data from other websites to prepare the questionnaire. All were hired as a team in Pangaea X.

Implementation

The team worked on 3 different things. The market researcher prepared a report on the datasets which was extracted from many available data sites published for free. The data scraping expert extracted a lot of information on customer’s questions occurring often. The data scientist then utilised these data points to create a model as training them with the questions which are commonly asked. The model was then deployed and tested real time and the efficiency was found to be 68% accurate.

Benefits Achieved

• Model achieved 68% accuracy
• All data procurement needs met in one single team