we are working as Finastra consultants with partners for a few years now and till now we haven't come across any solution for the final end-user. The volume of end-user is quite huge just for the sake of example total customers of just bank of America in 2009 were over 53 million out of which 30 million were online banking customers now just imagine the number of all the banks in 2019 This concept targets that the audience will save lots of end-users time and make their life easier. For the sales pitch point of Finastra, you can sell this product as it can cut the cost of banks for customer support representatives.
What it does
The customer can inquire about the account related details such as balance and can do financial transactions such as Internal Transfer with this FinBOT.
How we built it
We first built the training dataset for the model training that can learn from it and after training, it can predict future predictions, This is built using Tensorflow, Keras and layer dense such as bidirectional LSTM. The lemmatization is performed on text to normalize the data.
Challenges we ran into
There we multiple changes faced, while development in a very short span of time, The most challenging task is to prepare training data to collect financial industry datasets, and another challenge is to optimize the model to get maximum accuracy.
Accomplishments that we're proud of
We are excited, that by putting extra efforts we developed the Financial Bot that can save money and time.
What we learned
We learned in this Hackathon journey, how to do teamwork when you're running with a busy schedule and second the technical part, the model optimization was sucked us but after working / optimization/tuning we learned some new technique that helps a lot.
What's next for FinBOT
It's a FinBOT, our vision is to FinAll (Less Human Interaction In Financial Industry)