Advances in AI/ML automation technologies have been accelerating innovations in human-computer interaction (HCI) to improve overall efficiency and productivity while reducing human effort, especially for complex analytics tasks. Analysts have been assessing these technologies’ impact using subjective and qualitative surveys, but there has not been a systemic tool that automatically measures efficacy as analysts accomplish their tasks. This paper introduces SENSE, which attempts to measure human performance, behavior, and cognitive states in an automated fashion. SENSE is a tool-agnostic evaluation system to collect usability and utility metrics in real-world settings. SENSE consists of three components: (1) Analyst Testbed for data collection from neurophysiological sensors and behavioral inputs, (2) Backend Computational Models for performance, behavior, and cognitive state assessment for sensemaking and workflow optimization, and (3) Dashboard for administering data collection and conducting post-hoc analysis. We evaluated SENSE on Geospatial Intelligence (GEOINT) workflows with external evaluators, and SENSE was able to capture performance and behavioral inputs with 99% accuracy. SENSE also captured cognitive workload with strong correlation ( \(r=0.82\) ).