Every meeting that could be an email is a meeting that costs you a day of deep work. But async-first culture is harder than it sounds — until AI fills the communication gaps that forced you into synchronous rituals in the first place. This episode makes the case for a fundamentally different operating model.
Building an Async-First Company: How AI Replaces Synchronous Overhead
Yuki Tanaka runs Parallelwork, a fully distributed team of 14 across 9 time zones with a strict no-recurring-meetings policy. They use a custom AI layer built on top of Notion and Slack to handle the ambient communication that most teams default to synchronous standups for.
The core insight from Yuki's approach: most meetings exist to resolve ambiguity that could have been eliminated earlier with better documentation and faster async feedback loops. AI dramatically compresses both — generating documentation drafts automatically and enabling near-instant async responses via trained context windows.
We get into the specific workflows Parallelwork uses: daily AI-generated digest emails summarizing channel activity, a bot that writes decision memos from Loom recordings, and an automated weekly report that replaces the all-hands. Yuki also shares the failure modes — the patterns that looked async on the surface but were actually just delayed synchronous communication.
Parallelwork's template library is available at parallelwork.io/templates. Related FounderBrief coverage at founderbrief.xyz/category/founder-leverage.