When interacting with a chatbot, there are not indications of what to say or how to say it. Without affordances, it’s difficult to know what to do at first.
Affordances are important. For example, you may have never interacted with a door before but you can usually figure out how it works. A flat panel makes it clear you should push while a handle indicates that you should pull.
Because large language models are reliant on text inputs (and text is universal and infinite), this presents a UX challenge. How do you indicate what the bot can and can’t do? How do you help them get the best answers without needing to understand it deeply?
Read Why Chatbots Are Not the Future.