Marketers see AI benefits in creative work, but trust issues slow adoption of autonomous agents
Generative AI tools embedded in marketing workflows are delivering measurable results - faster creative production, better KPI analysis, and streamlined asset creation. Yet adoption of more advanced AI agents and automation remains stuck at just under half of marketing teams, held back primarily by trust concerns and technical complexity.
The findings come from a fourth-quarter 2025 survey of 142 brand and agency professionals, plus interviews with marketing and technology executives overseeing AI investments.
Where generative AI is working
Generative AI sees higher adoption than predictive AI across marketing workflows. Eighty-two percent of respondents use it for creative production, 81% for marketing tasks, and 75% for external communications.
Unilever's Beauty AI studio, built with marketing services group Brandtech, illustrates the practical impact. The in-house system generates assets for paid social, display, and e-commerce. Before the system, Unilever created roughly 20 assets per campaign. Now it generates 400 per product, according to Selina Sykes, global VP and head of marketing transformation for beauty and wellbeing at Unilever.
"It's a different way of working," Sykes said. "We used to send briefs off and get content back. Now it's this agile, iterative approach."
Predictive AI sees less adoption overall but finds its strongest use in measurement. Forty-eight percent of respondents use it for KPI analysis. Kroger Precision Marketing, the data science and retail media division of Kroger, recently launched an AI-generated email digest for suppliers that analyzes performance across short and long-term KPIs.
Adoption lags for media buying and financial work
Marketing teams use both generative and predictive AI less frequently for media buying and planning. Only 35% use predictive AI for this work, and 25% use generative AI.
Some executives see untapped potential here. At the Digiday Media Buying Summit in October, agency professionals discussed using AI to analyze requests for proposal, offering a faster initial review before human teams take over. For smaller brands and agencies with limited staff, AI could handle routine optimization and A/B testing that larger firms automate through platform tools.
Agentic AI faces a trust problem
Agentic AI - systems that anticipate needs and execute tasks autonomously without constant human input - remains the least adopted form of AI among marketers. Fifty-four percent of survey respondents said their companies don't use it at all.
The core issue is trust. Unlike generative or predictive AI, agentic systems run without human feedback at each step. If an agent hallucinates or misinterprets information, downstream tasks compound the error. Marc Maleh, global CTO at design and technology agency Huge, said marketers need clear governance frameworks before delegating authority to autonomous agents.
"There is a governance concern, especially with agentic and data access," Maleh said. "What do I want to give an agentic AI the rights to do on my behalf?"
Eric Lee, CTO at creative technology agency Left Field Labs, said building trust happens incrementally. Teams start with small, defined tasks - what he calls "micromanaged agents" - before expanding their scope.
"Building up that trust is a process, just like it would be with a human," Lee said.
Technical complexity is the bigger barrier
Beyond trust, agentic AI's technical demands present a steeper obstacle. These systems typically link multiple tools - APIs, large language models, and other platforms - rather than functioning as standalone applications. That integration complexity means few vendors have gotten agentic AI to work reliably at scale.
Sarah Mehler, CEO of Left Field Labs, said the industry is shifting from collecting as much data as possible to determining the right amount needed for autonomous systems to function effectively.
"Agentic AI is really hard to do, and not everybody can," Mehler said. "There are very few examples of it actually working well in the world."
Matt Maher, founder of independent research and research firm M7 Innovations, said marketers need to prepare infrastructure now for agentic systems. That means ensuring websites and backend systems can work with autonomous agents, and stacking tools in ways that maximize efficiency across the marketing operation.
For AI for marketing teams, the practical takeaway is clear: generative and predictive AI deliver immediate value in creative and analytical work. Agentic AI remains a future capability - one worth planning for, but not yet ready for broad deployment.
Your membership also unlocks:
