The research methodology I used was "contextual inquiry". I observed the user performing tasks within the studied platform with an expert/apprentice relationship, using the context and avoiding intrusive questions.

To capture insights during the research process, I used the sequence model. This model describes in detail step by step what the user is doing including the intentions, triggers, breakdowns and workarounds.

Being a niche product, it was crucial to find a user with experience in this area. I chose a prospect who has experience conducting and analyzing discovery interviews in a corporate and startup environment. The user is a student in his 20s who interned as a product manager in the summer of 2021 and experienced the pain of analyzing interviews manually.

• Log in and upload the recording of your last interview.

• Analyze and export insights from your recording.

The user got stuck in the second task when before analyzing the file the program asked him "What is the interviewer?". The user claimed that with the information presented it was impossible to determine with certainty who the interviewer is and that he would need to communicate with the person who conducted the interview in the first place.

At the end of the analysis, the user could not export the result because the interface was confusing. The user claimed that the button seemed to be part of the frame and therefore did not click it.

Reading the first few sentences of the transcript should be enough to determine who the interviewer is, but this is not obvious to the user. The proposal is to have a popup with the first sentences of the transcription that appears right after clicking on "Auto Analyze".

A new "Export Insights" button that skips the step of going to another screen to export them.