MediaScience's AI Tool Lets Marketers Test Ad Elements One at a Time
MediaScience announced a new AI system that recreates advertisements so precisely that test audiences cannot tell the copies from originals. The technology, called Creative Twin, uses AI for Marketing to let researchers modify individual elements of an ad-a celebrity's appearance, a product color, a model's hairstyle-and measure exactly how each change affects viewer response.
The company tested the approach with 812 U.S. respondents in partnership with the Ehrenberg-Bass Institute. Audiences could not distinguish between original ads and AI-generated versions.
What the Technology Does
Once an ad exists in digital form, every component can be systematically changed. A shampoo commercial featuring a model with straight hair can be modified to show the same model with curly hair. A puppy food ad can swap a Labrador for a poodle. A coffee commercial can test whether a celebrity drives results or if a generic model performs equally well.
Each modified version retains the production quality of the original.
MediaScience tested the curly-hair variant in the shampoo example. Women who saw the AI-modified version with curly hair showed higher brand recognition, brand attitude, and purchase intent than those who saw the original straight-hair version.
Practical Applications for Marketers
The system addresses a longstanding problem: isolating which creative decisions actually move the needle. Casting choices, visual elements, and messaging typically get bundled into a single ad. Testing them separately has been expensive or impossible.
Creative Twin also enables personalization at scale. Instead of producing multiple ad versions, a brand can create one high-quality execution and digitally adapt it for different audience segments. A shampoo brand targets curly-haired women with one variant and straight-haired women with another-both from a single production.
Duane Varan, CEO of MediaScience, said the technology provides "unprecedented clarity about what truly drives effectiveness" and lets marketers "optimize and personalize ads without compromising on production quality."
For marketing managers overseeing creative budgets, the approach offers measurable returns on specific production decisions. It also reduces waste by eliminating the need to shoot multiple ad variations.
The Technical Foundation
MediaScience built Creative Twin using MediaPET.ai, a company spinoff that handles the AI work. The system generates replicas indistinguishable enough that human viewers cannot detect the difference.
The company presented the technology at the Advertising Research Foundation's Audience x Science conference in March 2026.
Your membership also unlocks:
