Digital Agent enhancement
JM Studios/Sitel Group
The Challenge
I was brought on to as a UX Researcher for a legacy project looking to enhance its "digital agent"/ Chatbot for it's e-commerce eyewear website. The purpose was to find if the Digital agent would correctly assist the user and direct them to the correct pages in the website.
Methodologies
User Interviews
User Testing
the Process
Upon starting the process, I made sure to have all the up to date data from the PM: Interview guide, research objective, participant screener and participation reward information. The PM would set up the interviews from a pool of pre-selected users and would hand them off to me at the beginning of each interview through a recorded Zoom call. Introducing myself to the user, I would then explain the purpose of the interview, direct them to a beta site prepared for them to test, and proceeded to give them a brief description of the scenario they were to navigate; careful to indicate that there were no right or wrong answers and to act as they would if they were navigating this website on their own, think aloud. The interview/user testing would take approximately 45 minutes to an hour. I completed 8 user interviews/tests, in which I was able to gather data and pinpoint recurring problems.
THE Results
Once done with all the user interviews/tests, I went back through the recorded Zoom calls to verify my data points and make sure I did not miss any other relevant information. My data showed how unsatisfied, frustrated and unhappy the users were left feeling when interacting with the Digital agent. The Digital agent failed to complete certain dictated tasks due to its limited NLP. Users would try a variation of commands to try to complete the task but they were unsuccessful. With my recorded data, I proceeded to share my findings with the Development team, letting them know the most recurring problems and bugs. With this, the Development team were able to fix the Digital agent and it is now in production.