Act, Confirm, or Stop? Smarter behavior for AI assistants, wearables & robots — Amit Desai, Roku

Voice AI systems often struggle with error-prone interactions, creating a significant bottleneck for user satisfaction as technology shifts toward embodied AI. Instead of focusing exclusively on increasing technical accuracy, developers should leverage a second optimization knob: intelligent conversational behavior. By implementing the Outcome User Cost Heuristic (OUCH), systems can minimize the "pain" of negative experiences—such as playing the wrong song or failing to understand a request—by assigning quantitative costs to specific outcomes. This framework allows for strategic decision-making, such as choosing when to reject a request or ask for confirmation, to optimize the overall user experience. This approach remains scalable across diverse interfaces, including smart speakers and televisions, by adjusting cost parameters to reflect the specific modalities and user effort required for each platform.