fiserv 2026
ux research repository
the question
why wasn't the team's research repository actually getting used?
the approach
interviews with designers, researchers, PMs, and leadership surfaced a clear pattern: people were asking a person for research, not searching for it. i designed and built a solution: an ai-assisted dashboard, orchestrated in python on top of the existing repository, that lets employees search and interact with research directly, save studies, and route ai-generated summaries through a researcher approval step before anything goes out.
my work is under an nda, so this page covers the general approach and interface only.
what i did
→ ran and synthesized 16 interviews across research, design, product, and leadership to diagnose why the repository wasn't being used
→ designed and built an ai-assisted research dashboard, including search, a saved-studies workflow, and a researcher approval step for ai summaries
→ presented findings and recommendations to the ux team and leadership
takeaways
what i learned building something end to end, not just researching it:
1.) research and design aren't separate jobs. building the thing my own research pointed to taught me more about interpretation gaps than any interview did.
2.) a repository is not a strategy. the biggest fix wasn't technical, it was making research something people trusted enough to actually go looking for.