Turn a founder’s judgment into a shared system.

A grant program with money going out in installments to newsrooms that had mostly been built without operational infrastructure. Every judgment call routed to one person. The standard for output was high, personally held, and not written down.

Two things: a program, and the product it ran on.

The program: definition-of-done checklists that each ended in a system status change. Timeliness tables on every handoff. Escalation thresholds with named owners. A review cadence. Four diagnostic pillars, reassessments roughly every six months, and a mentor structure with capped hours. I seeded mentors into the mandatory workshops so newsrooms could watch them work and self-select — which turned a matching problem into a marketplace with a warm discovery layer.

The product: I product-managed the assessment platform. Near-daily vendor contact, user testing, copy, and a feedback pipeline that attached three things to every issue — what it was, how urgent, and a proposed fix — while leaving the implementation call to the tech team.

Two cohorts ran with defined stages and named decision-makers. The platform shipped: assessment, report, focus-area recommendations, mentor matching with full directory override, and calendar booking, wired into monday.com so status changes notified the right people automatically.

The founder built the assessment instrument and had the platform idea. I product-managed it. The engineering was the tech team's.

As of September 2026: the quality standards were built and presented, and I can't confirm they were adopted. The platform shipped publicly before it was ready, and the workarounds became permanent. A later version was scoped after I left; I can't confirm it shipped.