How to (Build and) Test Conversational Interfaces
|Speaking from experience, I can tell you it’s virtually impossible to test intent. By this, I mean that when you are building a mostly spoken UI, such as for an Alexa skill, it is imperative to test your conversations early and often, but the kicker is that it can be really difficult to do with more traditional user testing techniques. Can you set up remote user testing sessions when there is no tangible thing or site to interact with? How do you perform in-person tests that mimic the “real experience” with lo-fi prototypes?
Having spent the better part of the last 2+ years researching, prototyping, testing, and building experiences for a multi-modal social robot and enterprise chatbots, I plan to share what worked well, including specific tools, techniques, and tips for success.