The Challenge
What Ozzie Was Facing
The founder had a narrow window. The AI assistant space was moving fast and the specific workflow niche he was targeting — async meeting preparation for distributed teams — would be copied quickly once someone proved it worked. He needed an MVP in market collecting real user behaviour before committing to a full product build. Speed was not a preference; it was survival.
The Solution
What We Built
We ran a 3-day scoping session and agreed on a ruthless constraint: the MVP would do exactly one thing well — generate a structured pre-meeting brief from a calendar invite and past meeting notes. No voice features, no integrations beyond Google Calendar, no custom templates. We shipped in 6 weeks and seeded it with 40 beta users. We built instrumentation first so every session generated usable learning.
Results