Audemic Insights onboarding — Show the value before making the ask
Audemic Insights is a mobile app: a 24/7 junior analyst for the intellectually curious that finds and extracts key insights from research papers and reports, and turns them into audio summaries personalized to your goals. Retention was the problem. Users came for a seamless AI experience, met a form, and left before the app had shown them anything.
Role: Senior Digital Product Manager
Contribution: Product design, User interviews, Funnel analysis, Prototyping, Usability testing
Users expected a seamless AI experience and met manual tasks instead. Search results lacked precision. And an excessive number of onboarding steps produced friction that led straight to drop-off.
Linked the first search to topics worth engaging with, and added metadata that makes a result actionable rather than just present.
Onboarding leads with search, so someone explores a relevant topic before being asked for anything, and five personalized summaries show what the platform is for. Value first, then the ask.
An Oxford AI professor, a White House Correspondent, and a vaccine researcher highlighted the importance of showcasing value before making an ask from the user, and revealed that Google Scholar lacked 'real world' search scenarios like learning methodologies.
Mix panel analysis revealed significant drop-offs during sub-topic searches, pointing to gaps in relevance and user alignment.
Sign-up at 100%, the first interest question at 37.8%, continue-onboarding at 36.59%, sub-topic selection at 18.29%. Meanwhile 38.49% skipped onboarding entirely. More people skipping the flow than finishing it isn't a copy problem.
We hypothesized that advanced filters, prioritizing search, and enriching metadata in summaries would better meet user needs. To validate, we used an iterative approach with cycles of testing and refinement.