Managing interruptions effectively is a key challenge in modern work environments, as poorly timed disruptions during periods of high cognitive load can severely reduce team productivity and individual focus. While Intelligent Interruption Management Systems aim to solve this by assessing user states, they often face a critical trade-off between the precision of physiological sensing and the imperative of user privacy. We introduce CL-PPIIMS, a system built on a privacy-preserving architecture that resolves this conflict by applying Soft Computing principles—specifically, classifying imprecise cognitive states from noisy physiological data—to enhance Human-Computer Interaction. It leverages a stationary eye tracker for high-fidelity cognitive load data but performs all sensitive computations strictly on the user’s local machine. Only a highly abstracted, non-identifiable classification of the user’s state is then transmitted via Bluetooth Low Energy (BLE) to a mobile client, enabling an intelligent, privacy-aware management of team-based communication requests that enhances productivity without compromising user data sovereignty.