Designing AI that feels human: conversation design in the generative AI era

In this episode of AI Activations, Cobus Greyling sits down with Soo-jin Lee, AI conversational designer and personality architect, whose work has shaped AI experiences used by hundreds of millions of people across companies like Amazon, Google, Netflix, and Wells Fargo.

Soo-jin shares how her background in teaching and playwriting influences the way she designs conversational AI. The discussion explores why many AI systems fail to connect with users, how brands should think about tone and personality, and why “user patience” is one of the most overlooked elements in conversational design.

The conversation also dives into the shift from deterministic chatbots to generative AI agents, and how the role of conversation designers is evolving from writing every line to architecting experiences and guiding models.

Topics include:

• Why AI experiences must be designed for the broadest audience
• The importance of empathy in customer service AI
• Brand voice vs. user experience in conversational systems
• Designing for failure paths and real-world edge cases
• How conversation designers are evolving in the GenAI era