The growing integration of virtual assistants in vehicles raises concerns about their effect on driver stress and distraction. This study examines whether such an assistant can reduce stress or introduce risks. Using a driving simulator, we created scenarios with stress-inducing events such as pedestrian crossings, speed cameras, low fuel warnings, and accidents. The assistant intervened by providing reassurance, adjusting routes, and delivering non-driving-related information. The qualitative results show both benefits and drawbacks. Reassurance functions, such as addressing speed camera flashes or traffic congestion, effectively reduced stress. However, non-driving-related information, including tourist details and reminders, was often perceived as intrusive, especially under time constraints. A generational divide emerged, with older drivers preferring a strictly driving-focused assistant, while younger drivers saw it as a multitasking tool for both driving and contextual interactions. Participants also emphasized the need for greater personalization, allowing the assistant to adapt to their experience level, route familiarity, and stress levels. These findings highlight the importance of a context-aware and adaptive assistant that prioritizes relevant interactions while minimizing distractions. Future research should explore AI driven personalization, real-time stress detection, and user-controlled settings to enhance its integration into the driving experience, ensuring both safety and user satisfaction.