Despite the growing adoption of AI in finance, a critical gap persists between algorithmic prediction and actionable, ethical interpretation for investors. This study addresses this gap by designing and evaluating a novel hybrid AI-driven financial advisory system that seamlessly integrates predictive modeling (Gradient Boosting Machines [GBM], generative AI (GPT-4) and human oversight (Advisor-in-the-Loop) to deliver personalized, interpretable and ethically sound investment recommendations.
Guided by a design science research methodology, we developed an artifact that clusters 1,542 investors using K-means, predicts optimal investment strategies via GBM and translates these predictions into natural language advice using GPT-4. A key innovation is the embedded AITL framework, where financial experts validate all AI-generated outputs to ensure ethical and contextual relevance.
The hybrid system demonstrates high predictive accuracy (R2 = 0.85) and successfully segments investors into distinct risk profiles. More importantly, it bridges the interpretability chasm by converting complex model outputs into plain-language narratives. The AITL mechanism ensures the recommendations are not only accurate but also trustworthy and aligned with individual investor contexts, thereby promoting financial inclusion.
The system has not yet deployed in real-world settings. Future work should pilot-test the system across institutions and explore cross-market adaptability, dynamic data handling and ethical considerations such as AI fairness and explainability.
Although the system utilizes real-world data, it has not yet implemented in practical, real-world settings. Future research should focus on piloting the system across different financial institutions and evaluating its adaptability across diverse markets, its ability to manage dynamic data and ethical considerations such as fairness and the explainability of Generative AI (GAI).
The system provides personalized and plain-language guidance that enables informed decision-making for novices and everyday investors. It contributes to the advancement of financial decision-making in the context of investment.
This paper is among the first to integrate GBM and GPT-4 within a Design Science framework for financial decision support. Its primary contribution is the novel AITL model, which offers a blueprint for responsible, human-centered AI in management by effectively closing the loop between algorithmic prediction and managerial interpretation.
