Normalized claim
Quantified impact: 5 days
The workflow from planning to distribution now takes 5 days instead of months.
Dentsu Digital, a Japan-based digital marketing company, uses Amazon Nova Foundation Models through Amazon Bedrock in its integrated creative-production engine, Mugen AI. The solution supports ad-creative production, keyword identification, metadata extraction, multi-image analysis, performance prediction, and image and video generation for banners and ads. The company says the workflow from planning to distribution has been reduced from months to 5 days, and one client saw impressions up 255%, clicks up 367%, conversion rate up 861%, and cost per acquisition down 73%.
Reported outcomes
+861%
costCost savings
Strategic outcomes
Catalog median for cost savings deployments: +53.5% across 16 reported metrics. Compare benchmarks →
Normalized claim
Quantified impact: 5 days
The workflow from planning to distribution now takes 5 days instead of months.
Normalized claim
Quantified impact: 255% increase
One client saw a 255% increase in impressions and 367% increase in clicks.
Normalized claim
Quantified impact: 367% increase
One client saw a 255% increase in impressions and 367% increase in clicks.
Normalized claim
Cost: 861% increase
The same campaign saw banner conversion rate increase by 861% and cost per acquisition fall by 73%.
Normalized claim
Cost: 73% increase
The same campaign saw banner conversion rate increase by 861% and cost per acquisition fall by 73%.
No explicit deployment-stage evidence found.
Primary read
Showing 3 of 4
Dentsu Digital's Mugen AI creative-production engine uses Amazon Nova Foundation Models through Amazon Bedrock. Nova Pro interprets prompts, performs metadata extraction and multi-image analysis, and provides creative suggestions based on ad distribution data. It then routes tasks to Amazon Nova Canvas for image generation or Amazon Nova Reel for video generation. The solution is also positioned to expand into Amazon Nova Sonic and future multi-agent AI systems.
AI-generated summary. Verify important details with the linked sources before relying on this case.
Was this useful?
Community
No published comments yet.