Totally agree! I learned many of these lessons the hard way when we implemented AI chat bots in the LinkedIn experience. While the chat bot might be capable of doing a lot for the user, the users don't know that upfront. You have to make it easy for users to know what they are signing up for
Whether it's a chatbot or a contextual button, the question still remains: did the customer actually succeed / get what they wanted? Most teams building AI features today can answer 'did the user engage?' but not the 'did they accomplish their goal?' That's the real blind spot, regardless of UX paradigm.
Agreed, but before we can think about that we need users to try these AI features. The percentage of users not even trying them are disturbingly high...
"Do it for me" is the trigger, and not necessarily the UX required. I feel like there will be a set of users who want "tell me the next step" the first time the AI does it for them, and others who will say YOLO.
Ironically, Clippy or equivalent is an interesting UX model to explore here. "It looks like you're trying to ... do you want me to do it for you?"
I can't wait for you to see how I've approached UI/UX with the Par Not Far app. If we are so smart, why are we not helping with suggestions and intent mappers that can help guide the conversation. As brilliant as ML is, the empty text box and microphone icons are incredibly dumb UI for so much capability.
Totally agree! I learned many of these lessons the hard way when we implemented AI chat bots in the LinkedIn experience. While the chat bot might be capable of doing a lot for the user, the users don't know that upfront. You have to make it easy for users to know what they are signing up for
Whether it's a chatbot or a contextual button, the question still remains: did the customer actually succeed / get what they wanted? Most teams building AI features today can answer 'did the user engage?' but not the 'did they accomplish their goal?' That's the real blind spot, regardless of UX paradigm.
Agreed, but before we can think about that we need users to try these AI features. The percentage of users not even trying them are disturbingly high...
"Do it for me" is the trigger, and not necessarily the UX required. I feel like there will be a set of users who want "tell me the next step" the first time the AI does it for them, and others who will say YOLO.
Ironically, Clippy or equivalent is an interesting UX model to explore here. "It looks like you're trying to ... do you want me to do it for you?"
It's both, depends on the context! My todo list app offers to "Do it for me" if it can based on the content of the todo.
I can't wait for you to see how I've approached UI/UX with the Par Not Far app. If we are so smart, why are we not helping with suggestions and intent mappers that can help guide the conversation. As brilliant as ML is, the empty text box and microphone icons are incredibly dumb UI for so much capability.