The introduction of enterprise Generative AI (GAI) solutions, like Microsoft 365 Copilot, is transforming workplace practices. But, while GAI promises major efficiency and performance gains, it is not clear what the dynamics and outcomes of humans interacting with GAI in the workplace are. Drawing on 40 semi-structured interviews of a pilot project in a private organization, this paper explores how employees navigate the dual experiences of techno distress and techno eustress triggered by GAI. Using Grounded Theory Methods, I theorize three key mechanisms behind these tensions: GAI’s unknown capabilities, its fast evolution, and its diverse, sometimes unintended, uses. I contribute to the Information Systems literature on technostress by identifying new drivers of techno eustress and expand Human-AI Interaction literature by uncovering the dynamics of prolonged human engagement with GAI.