Easy Assistant: An App Design Study to Alleviate Daily Work Anxiety Among College Faculty

INTRODUCTION: Faculty juggle teaching, research, administration, and student support amid unstructured email, chat, and LMS alerts. These demands are linked to increased extraneous cognitive load and are associated with work anxiety. We evaluate a CLT-guided NLP app that structures messages into tasks. OBJECTIVES: Assess technical performance and preliminary pre–post changes in anxiety and workload. METHODS: Stage I validated a four-stage pipeline (cleaning; intent/entity recognition; probability calibration (evaluated by ECE); task structuring with priority gating). Stage II was a 4-week single-group pre–post feasibility study (N = 30) using GAD-7 and NASA-TLX. RESULTS: Intent F1 = 0.92; entity F1 = 0.89; 87% of tasks met TC ≥ 0.8. GAD-7 decreased by 2.7 points (95% CI [−3.6, −1.8]) and NASA-TLX by 16.3 (95% CI [−20.5, −12.1]); usability was high (SUS = 78.2). CONCLUSION: We observed pre–post reductions in workload and anxiety after four weeks of use; causal inference awaits controlled trials.