AI-POWERED BEHAVIORAL ROBO-ADVISOR FOR PERSONAL FINANCE DECISIONS

Abstract

There has been rapid growth in retail investors in India in the past few years. However, the biases that influence financial decisions are not being addressed by existing advisory tools. In this project, we present a behavioural robo-advisor driven by Artificial Intelligence. It is implemented as an interactive web application using Python and Streamlit. The system uses survey responses from 135 Indian investors spread across 5 age categories and sector allocation data of 220 brokerage client profiles to provide behaviourally informed advice. A survey questionnaire helps estimate the investor's exposure to each of the 11 most prominent behavioral biases and provide both a Behavioural Finance score and Risk Appetite classification. We have used a Decision Tree Classifier trained on the brokerage client data to estimate probable sector allocation based on age and gender. Further, we developed a module to perform a portfolio analysis of an uploaded set of equity holdings and detect the portfolio pattern which conforms with the investor's dominant bias. This report then identifies specific behaviorally driven insights based on actual portfolio behavior, avoiding offering any investment suggestions. Our approach aims to bridge the theory of behavioral finance with practical application while enabling development of scalable tools that are behaviorally more responsible.