Your AI isn’t producing bad creative, your brief is

The report by WARC, TikTok, and LIONS Advisory found that while 88% of marketers are producing more creative content since adopting AI, only 45% believe the quality has improved. The issue lies not in the technology but in the brief provided to generative models. Two-thirds of marketers brief models using demographic data, which they acknowledge is no longer effective.

Only 17% consistently brief with richer information like community context or behavioral signals. The evaluation layer that used to catch weaknesses in the brief has been removed due to time and cost constraints. Teams now brief with data they don't believe in simply because it's already available in planning decks, media plans, and audience field information.

This laziness in briefing leads to plausible but flawed creative outputs that go unnoticed. The compression of timelines in Southeast Asia, where teams manage multiple markets with limited budgets, exacerbates the issue. The regions with tighter budgets and more markets struggle to incorporate nuanced local insights, resulting in generic content that fails to resonate.

Teams that treat the brief as a product, investing in tools that generate audience insights and validate content performance, are seeing better results. This shift in approach could help rebuild the evaluation layer and improve the quality of AI-generated creative content at a relatively low cost.