Why Nobody Wants to Learn Your Chatbot

The challenge

The organization wanted to introduce a new experience that taught customers what its support chatbot could do, with the goal of increasing engagement and reducing support calls. The question wasn't whether the content was useful, but whether people would actually choose to engage with it. That distinction was especially important in a support context, where users are likely arriving in a state of frustration, not exploration.

Approach

I conducted moderated interviews with both current and prospective customers, exploring how they approached support, interacted with chatbots, and reacted to a prototype designed to showcase the assistant's capabilities.

Key insights

Participants did finish the experience with significantly greater confidence in the chatbot's abilities. *However*, despite their positive reactions, users had little motivation to voluntarily explore educational content.

Key learning: People arrive at support experiences with a goal to solve a problem, not to learn a new tool. Any learning needs to be automatic, lightweight, and easy to skip.

Impact

The findings shifted the conversation from improving educational content to rethinking onboarding altogether. Rather than asking users to invest effort at a time that they were already in need of support, the team began reevaluating the first-time experience to include key capabilities naturally as part of the support journey.