video1.0<iframe src="https://www.loom.com/embed/9e2842def0c04acd85b9d1572cd20a43" frameborder="0" width="1108" height="831" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>8311108Loomhttps://www.loom.com8311108https://cdn.loom.com/sessions/thumbnails/9e2842def0c04acd85b9d1572cd20a43-def4553c1ba062ee.gif369.067This Loom reviews how a weekly research agent drove feature improvements for insight sharing and a stronger startup activation flow. The author updated the onboarding and entry experience so users can generate a populated study by entering a URL, then preview and run it with attached files and contextual setup. They tested the experience with 10 startups, noting iterative implementation effort, adding an extra user step because context was insufficient for the agent to generate good plans. While the alpha is close and working well, they still plan refinements such as fixing issues with end dashes and correcting a goodwill model placement.