In the Age of AI, Creative Taste Is the Last True Moat

Summary

The MAU 2026 panel moderated by Liftoff's Casie Jordan, featuring Singular's Stephanie Pilon and Underdog's Kaylee Wilmovsky, addressed a central tension in AI-driven mobile creative: whether teams are actually improving output or simply producing more of it.

Key data point: Singular's upcoming report indicates top quartile advertisers launch more than 50 creative variations per week, with the upper range operating 5,000-10,000 active creatives simultaneously. Creative volume correlates with spend, creating structural disadvantages for lower-budget teams. Wilmovsky emphasized that running low-volume tests without adequate budget generates the feeling of rigor without the substance.

The Creative Director role is shifting from executor to orchestrator. AI manages the output layer (volume, iteration, platform-native variants), while humans provide strategic direction, taste, and editorial judgment. Hiring priorities are moving toward curiosity, critical thinking, and the capacity to recognize when AI elevates versus erodes distinctive creative work.

On "human-made" creative, panelists could not reach a clean definition given AI involvement in prompting, directing, and approving assets. The uncanny valley effect persists. The practical reframe: evaluate creative on whether it shows up natively for each platform's audience rather than whether AI produced it.

Stop doing: tests without budget or conviction. Double down on: systems that compound over time rather than quarterly campaign resets. The competitive advantage is increasingly tied to judgment, not tooling.

Analyst Note

This panel captures a transitional moment in mobile marketing creative production. The signal worth noting: AI is industrializing the output layer while leaving editorial judgment as the differentiating constraint. Competitive advantage is shifting from execution capacity to taste and direction.

What's notable is the convergence on a definition problem. "Human-made" becoming contested signals the industry has moved past the point where AI involvement can be cleanly bracketed out of creative workflows. The implication for UA teams: provenance is losing relevance as a strategic frame, while platform-native authenticity is gaining ground as the evaluative criterion.

For ad ops specifically, the structural correlation between spend and creative volume has been an open secret. Singular's data point that top quartile advertisers run 5,000-10,000 active creatives frames this as a scale problem, not merely a tooling gap. Worth watching: whether smaller-budget teams develop counterstrategies such as tighter hypothesis-driven testing or smarter iteration cycles, or whether the divide widens as AI tools compound advantages for high-spenders.

The broader trend context: as privacy regulations constrain targeting precision, creative is becoming the primary lever for performance differentiation. Teams treating AI as a volume multiplier without corresponding investment in judgment and taste are positioned to hit diminishing returns faster than peers who invest in the orchestrator function the panel describes.