SageSpark - Helping founders act on AI feedback

Early-stage founders struggle to turn a raw idea into a pitch-ready venture thesis. SageSpark built a multi-agent AI system for this, with specialized agents working together to guide each founder. When engagement stalled, I led the study as the solo UX researcher to find out why.

I triangulated 17 contextual interviews, 6 moderated usability sessions and a heuristic evaluation. Over 60% of founders abandoned the workflow mid-task. Three breakdowns drove it: open-ended prompts that made them freeze, dense AI outputs that overloaded them, and a plain-language barrier from heavy business jargon. Every founder in usability testing copied the content into Notion or Google Docs just to process it.

I synthesized the findings into personas and framed the core problem in one line: founders could not act on what the AI gave them. I then ran concept testing on an interaction model built on guided prompts, structured scaffolding and progressive disclosure.

Task efficiency improved 55%: time-on-task dropped from 10 minutes to 4.5. My findings reprioritized the six-month product roadmap.