Pivotal partnered with Bidstack to automate image evaluation and cross-platform ad detection using advanced computer vision models — reducing manual workload and unlocking new ways to measure media performance.
Game publishers were manually reviewing thousands of ad images to ensure brand safety and compliance, delaying campaign approvals for AAA game titles.
Meanwhile, existing measurement systems excluded ad views from streaming platforms like Twitch and YouTube, under-reporting exposure and lowering perceived ad value.
We deployed vision training models to classify ad content automatically, replacing manual review.
Pivotal Built models to identify and block inappropriate content — including adult, violent, or hateful imagery — using configurable confidence thresholds.
Created logic to locate and match ads within fast-changing video frames across gaming and streaming content.
Linked detections to impression and exposure data to calculate downstream ad value across in-game and streaming channels.

Eliminated manual review for AAA publishers, freeing teams from time-consuming workflows.
Expanded measurement into streaming for the first time, unlocking new insight beyond traditional media.
Improved analytical precision and significantly reduced exposure to brand safety risk.
Demonstrated the power of advanced AI in complex media environments, proving performance at enterprise scale.
We don't just advise, we roll up our sleeves and build alongside your team. We understand how to make products investors can believe in and customers actually use.