Investor suitability assessment remains a critical component of institutional onboarding processes, particularly within digital investment platforms offering access to complex and risk-sensitive financial instruments. Conventional profiling approaches based solely on static financial disclosures often fail to capture the behavioral dynamics that influence investment intent and decision-making under uncertainty. The present study proposes an explainable artificial intelligence (XAI)-driven framework that integrates structured financial indicators with interaction-derived behavioral signals to enhance suitability classification during onboarding. Financial variables such as annual income, net investable assets, liquidity ratio, and prior investment experience were combined with behavioral metrics including investment exploration latency, decision response time, asset browsing entropy, and portfolio reallocation intent to construct a multidimensional suitability assessment model. Supervised ensemble-based classification techniques were employed to segment investors into Conservative, Moderate, and Aggressive suitability tiers, while post-hoc explainability methods enabled the interpretation of feature-level contributions to classification outcomes. It is based primarily on a Random Forest classifier, selected for its robustness in handling nonlinear relationships, mixed data types, and high-dimensional behavioral–financial datasets. The results demonstrate improved predictive accuracy and interpretability through the integration of behavioral engagement patterns alongside traditional financial metrics. Distributional analysis and hierarchical clustering further revealed distinct investor groupings aligned with suitability tiers, thereby supporting the relevance of multidomain onboarding indicators in investor profiling. The proposed approach contributes to the development of transparent, adaptive, and compliance-aligned onboarding systems capable of minimizing suitability misclassification and enhancing investment-product alignment.
Neelesh Lalwani (Sat,) studied this question.