Skip to main content
Case Study · Media & Content Production

Regensys 10x’d media processing by augmenting 300+ editors with a custom AI pipeline

Linkt’s FDE team built a custom-trained model and pipeline that turned Regensys’s 300+ photo, image, and video editors into AI-augmented versions of themselves.

Regensys
Media & Content Production
FDEAgents
10x
Media processing throughput
300+
Editors augmented
3
Media types supported
Photo, image, and video
The Challenge

Output tied to headcount in a high-volume media operation

Regensys has a team of 300+ editors working across photo, image, and video. Output was linearly tied to headcount—every new client meant more editors, with no efficiency gain.

They needed to multiply their team’s output without replacing them. The solution had to work across all three media types and meet existing quality standards.

The Approach

A custom-trained model built around how editors actually work

Linkt’s FDE team embedded with Regensys’s editorial leadership to build AI tooling that amplified each editor’s skills rather than replacing them.

  1. 1The FDE team mapped editorial processes across photo, image, and video—identifying where AI assistance would create the most leverage.
  2. 2A custom model was trained on Regensys’s historical output, learning the company’s specific quality standards and editing patterns.
  3. 3The model was integrated into a production pipeline where AI handled initial processing and editors reviewed, refined, and approved output.
  4. 4A unified interface let editors move between all three media types without switching tools.
The Outcome

Same team, 10x the throughput

The 300+ editors went from manual processing to AI-augmented workflows, reaching up to 10x the previous throughput. The custom model matched Regensys’s quality standards because it was trained on their own work.

No additional editors were hired. The capacity increase came entirely from augmenting the existing team.

Technology Stack
Custom Model TrainingFDE EngagementMedia Processing PipelineMulti-format Support
Explore more impact studies