loder

BFSI

Clients

The customer is an US based FinTech firm, dedicated to providing a stock trading platform as a virtual assistant. It handles approximately 300K clients. The company was keen on adopting Machine Learning in their trading style, allowing the customer to feel that they are interacting with a trading expert or a broker, who showcases recommendations on which stocks to invest in. They have a commission based revenue model, that is whenever a customer gains a profit, they take a small commission out of the transactions.

img

Problem statement

The trading platform followed a process that involved severe amounts manual trading, making the process as a whole, very complex for the customers. Due to the dynamic nature of the stock market, the firm wanted an easier method of predicting/suggesting stocks for/to their customers.

Mitigating the problem

A hybrid AI solution was proposed which solved the problem

  • Talking to an AI bot which was trained specifically for stock-market trading purposes made their communication easy.
  • Automating the task of verifications and trading operators.
  • Reinforcement Learning and Time Series analysis was implemented on live streaming stock ticker data.
  • Recommendation system was setup which would analyse the behaviour of user in trading and would recommend to go for specific stock based on analysis.
img

Solution delivered

  • The bot integration gave customers a feeling that they were interacting with industry experts. This solved the problem of the users and it increased the number of users by 32% in past 2 quarters.
  • The stock market predictions and recommendation system showcased accuracy which pulled the market by 13% in past 2 quarters.
  • The ROI increased by 18% as automating the entire process let to quick decisions.