What changed, and what it took.

Scaling a pipeline to deliver a $2M annual contract

A nearly full pipeline scaled to deliver a $2M annual contract, lowering operating costs and enabling revenue recognition a quarter ahead of plan.

From two days to under ten minutes for case evaluations

A Slack workflow freed operations from an engineering queue, while user feedback shaped a prototype for the broader policy lifecycle.

Human approval retained on every case.

How customer research changed the platform roadmap

A build-or-buy question became a question about who the platform was for, and research shifted the roadmap from a social-network concept toward efficient content creation, customization, and curation.

Increasing adoption by preserving campaign control

Adoption research led to simpler automation that preserved campaign choices, eliminated daily budget adjustments, and increased adoption by 25 percentage points after the first release.

Unblocking a catalog migration with 94% automatic image matching

Automated reconciliation of approximately 20,000 artwork images unblocked the catalog migration onto ArtLifting's new ERP, turning a weeks-long manual audit into days.

Before you can trust the AI, find out if your humans agree

Evaluating existing wildlife tools and correcting reference labels led to a platform the sanctuary adopted to clear its image backlog.

Classifying four million CTV videos in 30 days

With no CTV models available and 30 days to deliver, the team used GPT-3.5 to classify four million additional videos and meet a deadline the other partners missed.

Adoption looked healthy. Two teams had quietly stopped trusting the data.

Two major client teams had routed around Looker and were collating reports by hand. Following the workaround instead of building another dashboard cut sixteen to twenty hours of manual collation per cycle to thirty minutes of QA, while protecting the insight work that was the actual value.