Explainable and Context-Aware Financial Nudges via Event- Driven Microservices

Abstract

As financial decision-making becomes increasingly automated and personalized, users demand not only timely and relevant financial recommendations, but also transparency in how those suggestions are generated. This paper presents a novel microservices-based framework that delivers context-aware financial nudges enhanced by explainable AI (XAI), designed for real-time deployment in modern fintech applications. The proposed system leverages event-driven microservices to continuously ingest and process multi-modal data streams including transaction history, geo location, time-of-day patterns, and behavioral signals to deliver actionable insights such as spending warnings, savings opportunities, and goal reminders. Each microservice is responsible for a modular task such as context classification, recommendation generation, or explanation rendering. What distinguishes this work is the integration of explainability modules using interpretable AI techniques (e.g., SHAP, counterfactuals, rule-based trees) embedded within each service. This enables the system to answer, "Why am I getting this nudge?" in natural language, thereby fostering user trust and behavioral compliance. The framework is evaluated on synthetic and anonymized financial datasets to simulate diverse user behaviors. Results demonstrate the effectiveness of contextual triggers (e.g., time, location, prior habits) in increasing user engagement, while explainability boosts users' perceived relevance and trust in the system. The architecture adheres to principles of modularity, fault isolation, and data minimization, making it suitable for deployment in privacy-sensitive financial environments. This research bridges the gap between intelligent personalization and transparent automation in fintech, paving the way for ethical, user-centered financial advisory systems