This research work explores the use of Artificial Intelligence (AI) in personalized portfolio management and its effectiveness in asset allocation, risk management, and decision-making for investors, besides exploring the possibility of Human-AI collaboration. The primary data was collected from 100 active individual investors. This research uses correlation analysis, linear regression, and chi-square tests to determine the investor preferences for AI-based portfolio management systems and human financial advisors on eight decision-making criteria: financial literacy, risk tolerance, investment horizon, liquidity, taxation, macroeconomic awareness, behavioural biases, and ESG preferences. The results show a clear demarcation of portfolio management variables based on their automatability. Financial literacy, risk tolerance, investment horizon, tax benefits, liquidity, and macroeconomic awareness have strong positive correlations with AI preference, indicating high automation possibilities for algorithmic systems. Behavioural biases and ESG preferences have weak or negative correlations, indicating limitations of AI in dealing with psychological and value-based investment issues. Regression analysis shows that while income positively affects investable capital, it explains only a small variation, thereby emphasizing the need for multi-factor profiling in personalized portfolio management. Chi-square tests show a statistically significant overall preference for human financial advisors, with most investors preferring hybrid Human-AI advisory models over fully automated systems. The findings of this research work provide empirical validation to the behavioural finance and fintech literature by showing that AI improves efficiency in asset allocation and risk management but cannot fully replace human judgment in behavioural and ethical aspects. This research work concludes that the future of personalized portfolio management lies in AI-based advisory systems that leverage the power of human and artificial intelligence together, rather than in fully automated systems