Don’t design a conversation, design FOR conversation.

If we want computers to communicate more like humans, then we can't design one-sided conversations. Instead, we have to imagine interactions as complex and spontaneous as human communication.

User research and testing could help us further explore multimodal interactions.

Unfortunately, due to time constraints, we were not able to devote time to in-depth user research and testing. We are just scratching the surface of a multimodal conversational interface and there’s room to explore scenarios such as error-states or emergencies.
We can push back against trends and paradigms.

We created a compelling concept without compromising our values of an equitable and sustainable future. We also intentionally shifted away from servant-master/gendered voice UI models. Regardless of the context, our designs should strive to promote a better world.