TRANSFORMING ASSET ALLOCATION THEORY THROUGH ROBO-ADVISORY AND ARTIFICIAL INTELLIGENCE

Abstract

This study synthesises the global literature on the transformation of traditional asset-allocation theory, particularly Modern Portfolio Theory (MPT), toward decision models based on Artificial Intelligence (AI) and Robo-Advisory. Using a Systematic Literature Review (SLR) with the PRISMA 2020 protocol, 46 Scopus-indexed articles (2016–2026) were analysed. The findings document a paradigm shift from static mean-variance optimisation toward Deep Reinforcement Learning (DRL), hybrid FNN–DRL architectures, and the integration of alternative data such as ESG sentiment, Vision-Language climate perception, and digital assets. The study proposes an emerging "Algorithmic Portfolio Theory" and underlines the need for Explainable AI to secure accountability and investor trust in increasingly autonomous asset management.