Systems
Data, playbooks and AI
Data shows what's happening. Rules flag what needs a call. Playbooks set the standard response. AI and automation make the work faster and cheaper to deliver.
AI-assisted assetsSample
47%
Simulated · September · 22% in January
Cost per asset
$462
Simulated · 25% lower than January ($612)
Hours saved, YTD
4,130
Simulated estimate · 740 in September
Asset Factory margin
29%
Target 45%. Savings aren't reaching the price yet
Sample data, not production activity. These AI-production figures are simulated to show how I would measure AI: by quality, speed and cost of delivery, so the time it frees goes back into strategy and better work. No AI tools are connected to this prototype.
AI share and unit cost
Violet: share of assets with AI in the workflow. Green dashed: cost per asset, indexed to January = 100.
Adoption by workflow step
Share of assets where AI handled the step
- Resizing & versioning · automation templates88%
- Creative tagging · creative tagging71%
- Concept & storyboard ideation · Gen-AI image + LLM62%
- Localization & voice · AI VO & translation54%
- Background & product compositing · Gen-AI image47%
- Final animation · Mostly manual12%
Roadmap
What the unit is piloting and rolling out
- Agentic brief refinement in the automation platformPilot on Pulse and AtlasIn pilot
- Creative tagging on every account7 of 10 accounts taggedRolling out
- Auto QA for safe zones and specsTargets the top QA failureScoping
- Dynamic creative for catalogsAtlas and NorthpeakPlanned