New research published in the journal Marketing Science shows that an AI system designed to optimize for both performance and brand alignment can beat human designers in live advertising campaigns - and the advantage held up 18 months later. The findings challenge the common assumption that generative AI is useful mainly for rapid ideation or aesthetic polish, not for delivering measurable results.
Researchers Remi Daviet of INSEAD and Yohei Nishimura of the Wisconsin School of Business built an active-learning engine that generates ad visuals, vets them against brand standards, and field-tests them to find high performers. In a live Instagram campaign for an outdoor activities company, the AI-generated portfolio recorded a mean click-through rate of 0.98%, compared with 0.65% for a professional human designer and 0.78% for an AI model optimized only on aesthetics. The researchers' analysis showed the system's creatives beat the human designer's batch in 99.59% of samples.
Performance held up over time
A follow-up test conducted 18 months later, during a high-stakes booking season and without retraining the model, confirmed the durability of the learned signals. The AI portfolio achieved a mean click-through rate of 3.38% versus 3.24% for the company's contemporary human-designed campaign, again with lower variance.
"Our findings challenge the common assumption that generative AI is useful mainly for rapid ideation or aesthetic polish," said Daviet. "When performance prediction and brand alignment are jointly optimized through active learning, the resulting visuals can deliver higher average returns and lower creative risk than traditional processes."
What the system does differently
The key difference is how the AI searches for creative options. Rather than generating images based on aesthetics alone, the system uses an active-learning engine to predict which visuals will perform well while staying on-brand. Marketers generated background images promoting nature exploration, which were vetted for brand standards and field-tested.
For marketing teams, the practical implication is that the bottleneck in AI-driven creative work has shifted. "Strategies that treat AI as a simple production tool risk missing both the performance gains and the consistency that a purpose-built search-and-alignment system can unlock," said Nishimura. As generative models continue to improve, the competitive advantage comes from intelligent exploration of creative possibilities under brand constraints.
The findings matter for digital advertisers, creative agencies, and marketing platforms that are evaluating how to deploy AI in campaign production. For marketers specifically, the research suggests that pairing generative tools with performance data and brand guardrails can produce creative that outperforms traditional processes - not just in cost or speed, but in engagement metrics that matter to the bottom line. That's a relevant consideration for anyone managing AI for Marketing efforts or evaluating AI for Marketing Managers training options.
Why this matters for marketers
The study offers a concrete benchmark: an AI system built for performance prediction beat human designers by roughly 50% in click-through rate in the initial test, and maintained an edge in a follow-up test 18 months later. For marketing teams, that means the question is no longer whether AI can produce usable creative, but how to build systems that search for high-performing visuals under brand constraints. Teams that treat AI as a production tool rather than a search-and-optimization system are likely to leave performance on the table.
Your membership also unlocks:
