Hybrid AI system for interpretable student policy guidance using rule-based reasoning and retrieval-augmented generation
TL;DR
A hybrid artificial intelligence system integrating rule-based reasoning with retrieval-augmented generation (RAG), coordinated by a hybrid integration layer that routes each query to the appropriate reasoning path, demonstrating a viable, open-source-deployable AI-assisted solution for policy guidance in developing-country contexts.
Abstract
Digital platforms in Nigerian tertiary institutions often lack interpretive capabilities for complex academic policies, leading to inconsistent guidance and administrative burdens. Existing e-portals primarily process transactions without explaining policies, while general chatbots fail to handle the conditional, exception-laden nature of institutional rules. Here we developed and evaluated a hybrid artificial intelligence system integrating rule-based reasoning with retrieval-augmented generation (RAG), coordinated by a hybrid integration layer that routes each query to the appropriate reasoning path. Evaluated across two Nigerian universities, the system achieved 89.7% accuracy, a mean response time of 2.34 seconds, and an estimated 41.9% reduction in routine staff queries, while system-staff interpretation consistency (86.0%) exceeded staff-staff consistency (76.0%). The hybrid path delivered high-confidence, low-latency deterministic responses, while the RAG path maintained broad coverage. The system enhances consistency in policy interpretation and improves student access to institutional regulations, demonstrating a viable, open-source-deployable AI-assisted solution for policy guidance in developing-country contexts.