I’ve seen many articles in blogs or even here, with people citing the use of ChatGPT in UX to help in the research process. However, despite this being completely valid and as far as we have today, I like to think about future perspectives for this technology and what it can help in UX.
The main contribution I see is to scale moderate usability tests to be unmoderated and in user interviews, streamlining the process but still managing to obtain great insights.
During a usability test where the participant is being recorded, I can see the AI being used to analyze facial and body expressions and then extract valuable information according to interactions with the interface or according to the questions.
According to the interactions that the user, with the expression analyzed, and the previous answers, AI could continue with questions adapted to the context, just as we do in a moderate test/interview.
Like the generation of dynamic questions, we can have a faster analysis of the answers given by users with AI.
Usability tests can be time-consuming to be analyzed in depth, but with AI we can turn this task much faster to be completed and can also scale them for more people without losing in quality, i.e., doing qualitative research with a large number of people but faster. AI could identify behavioral patterns, interactions that users had difficulties, and help identify the improvements to be implemented.
Although Jakob Nielsen in his article for NN/g said that we only need 5 people to discover most of the usability issues, and more than that is a waste of resources, when we put AI in the middle we’re not going to waste so much of the resources and maybe we can find more hidden issues, do more tests, and more often.
There are already some apps that collect interesting insights during the call, either for future agreed tasks, or behavioral observations between speech and others. I have already used these tools and see that they generally work well, with small slips. But I can see this same technology being used to track user interviews, where the main focus would be to analyze the user’s answers, perhaps even suggest to the interviewer questions to move further.
With cheaper processes, we can do more UX research!
Read the full article here