A Unified Fuzzy–Explainable AI Framework (FAS-XAI) for Customer Service Value Prediction and Strategic Decision-Making

This record corresponds to the accepted manuscript (post-print) of the following journal article:

“FAS-XAI: A Unified Fuzzy and Explainable AI Framework for Interpretable Decision-Making under Uncertainty”

This article introduces FAS-XAI, a unified methodological framework designed to support transparent and interpretable decision-making in complex, real-world scenarios characterized by uncertainty and incomplete information. The framework integrates fuzzy clustering to identify latent behavioral profiles, supervised learning models to estimate decision outcomes, and Explainable Artificial Intelligence (XAI) techniques (SHAP, LIME, and ELI5) to ensure both global and local interpretability.

The methodology is validated through a real-world B2B customer service use case based on RFID indicators (Recency, Frequency, Importance, and Duration), where fuzzy customer profiles and explainable XGBoost predictions are combined to assess attrition risk and support strategic decision-making. The study highlights the coherence, interpretability, and adaptability of FAS-XAI, positioning it as a general-purpose, human-centered framework applicable across domains such as business, education, physics, and industry.

The final published version is available at the publisher’s website:

https://doi.org/10.3390/ai7010003

This deposit is made for open access and dissemination purposes, in accordance with the publisher’s self-archiving policy.