We do and need banking everyday in our life and we see how complex the banking websites have become. No matter how intuitive the UI, a user has to put labour to get around many operations and it's a hassle. So, we thought to make it easier and created 'Teller Alfred' a chatbot for this very task.
What it does
Our project demo is in a baby stage. Teller Alfred is essentially a chat box, where user can ask questions regarding account balances, deposits and transfers and the bot tries to provide a legitimate solution or information using NLTK and the Capital One APIs and SDKs in back end.
How I built it
We built it using Python with NLTK and deployed it using Django using AWS services.
Challenges I ran into
NLP : To find and prepare set of queries that a user might have to ask the chat bot, we brainstormed. We had to select the best library that could do this task of working on the data and identifying entities in the minimum amount of time.
UI: The UI had to be kept minimal, and focused on the job.
Accomplishments that I'm proud of
A beautiful little bot 'Teller Alfred' that can grow into a very big advantage in the field of banking, saving lot's of time on the part of the customer as well as the tellers and bankers, reducing the traffic on the portals too.
What I learned
NLTK, Chatbots and lot's more from the great teamwork!
What's next for Teller Alfred
We have a strong desire to extend the power of Teller Alfred using superior libraries and AI. We hope we are successful.