Jan Erik GrauSimon MayerJannis Rene StreckerKimberly GarciaKenan Bektas2024-03-282024-03-282024-05-11https://www.alexandria.unisg.ch/handle/20.500.14171/11977910.1145/3613905.3651066This paper introduces a novel system to enhance the spatiotemporal alignment of human abilities in agent-based workflows. This optimization is realized through the application of Linked Data and Semantic Web technologies and the system makes use of gaze data and contextual information. The showcased prototype demonstrates the feasibility of implementing such a system, where we specifically emphasize the system’s ability to constrain the dissemination of privacy-relevant information.en-USHuman-Agent-CollaborationSolidKoreografeyePrivacySemantic WebGazeNotificationsInnovationGaze-based Opportunistic Privacy-preserving Human-Agent Collaborationconference contribution