An AI powered chatbot for CitiBank. It supports multilingual support for a more wider reachabilty. It provides accurate sugestion for any user's query at very quick rate.
CitiBot.mp4
The CitiBot has a very fast suggestion system built with Elasticsearch running in Docker container. The data from Elastic Search is servered to the frontend with a help of REST API made with Fastapi framework. The data that elasticsearch is feed with by a crawler programed in python language which collects all the data from webpage in citibank. The user's query is searched through the data collected and the best results are recommended back to user with whichever language the query was made with .
Since around two billion people are actively using Whatsapp.Citibot can also be acessed through Whatapps to have easy usage by the consumer.This was possible by integrating our REST API to Twilio platform.
Citibot-Whatsapp.mp4
With Citibot we can analysis the user based on the query they ask,in which language was the query asked ,how frequently is the query asked and many other analysis parameters.Analysis is possible by using Kibana ,with which we can create real time dashboard to monitor the query asked to Citibot and do suitable action to help the users more .
Here is a dashboard we created for Citibot
