Conversational AI (CAI) is set to play a key role in Augmented Communication Technologies (ACTs), but what forms will it take, and what advantages or costs will it bring? Over the past year, my lab has developed AI systems designed to generate augmented speaker sentences that seamlessly integrate with ongoing conversations. Our focus is on creating conversational aids that close temporal gaps, facilitating smoother dialogue. By building our socialBot from the ground up, we have tested key assumptions about AI’s ability to represent individuals in conversations. In this presentation, I will explore the fundamentals of conversation, the dynamics of augmented speaker interaction, and the core components of CAI-enhanced ACT. Additionally, I will share insights from our work, along with tools and criteria for evaluating the effectiveness of CAI in enhancing conversational exchanges.