Quorum

Meeting-notes automation for an India-based venture fund. Watches deal-stage changes and produces IC-memo drafts grounded in time-windowed evidence per stage.

Client NameIndia-based VC fund
ServiceAI workflow automation
Quorum AI workflow automation — meeting-notes-to-IC-memo pipeline triggered by ClickUp deal-stage changes.

What we built

Quorum is an internal automation we built for an India-based early-stage venture fund. It pulls meeting transcripts from four AI notetakers (Fireflies, Read.ai, Fathom, Granola), maps each conversation to a startup in the portfolio, and drafts structured due-diligence memos directly into the analyst's ClickUp workflow. It runs on a Sonnet 4.6 call with a system prompt that has gotten shorter over time, not longer.

What's interesting about it

The model wasn't the hard part. The trigger was. We tried four different ways to fire the agent — a cron job, contextual toggles, an explicit checkbox — before we found the one that produced near-100% adoption. The breakthrough was watching the work itself, not adding a feature.

Read the full engineering story

The four-trigger journey, what we got wrong, and why we now believe the trigger is the product is written up in detail. Four ways we tried to trigger an AI agent → only the fourth one worked.

Quorum's four trigger layers visualised — cron, toggles, checkbox, and the workflow-state-change trigger that finally worked.
Quorum project hero — a hand sliding a kanban card between stages, the moment that became the agent's trigger.
PactImagika