A new survey of 300 chief marketing officers finds almost every one calls AI transformative, but only 8% run campaigns where multiple AI agents operate autonomously, an 88 point gap between rhetoric and reality.
The distance between how chief marketing officers describe their AI ambitions and what they actually run has rarely been this wide. A survey of 300 CMOs and follow-on interviews with 50 of them, published by Boston Consulting Group and distributed June 15, 2026, found 96% of marketing leaders say artificial intelligence is driving an end-to-end transformation of their function. Only 8% are running campaigns in which multiple AI agents operate autonomously, a model the report calls "agentic marketing." The 88-point spread between rhetoric and execution is the story, and BCG frames it as the central gap separating marketing's leaders from its laggards.
Most of the chief marketing officers surveyed are not even close to the leading edge. About 42% use generative AI, or GenAI, the class of systems that produces text, images, and other media on demand, only as a task assistant in a handful of workflows, drafting copy here, summarizing a brief there. A third say they have transformed significant parts of their function with agents, but on closer reading that includes single-agent automation and tightly scoped tools, not the multi-agent orchestration the report treats as the actual finish line. Just under half of the respondent companies, 43%, report AI investments in marketing topped $15 million this year, a level of spend that should by rights have produced more operational change than the survey captures.
What separates the 8% from the 42%, the report argues, is architectural rather than budgetary. The chief marketing officers running agentic campaigns have typically built four things the others have not. The first is a unified data foundation, the customer, content, and performance signal a marketing organization actually owns and can feed into models without weeks of integration work. The second is a brand intelligence layer, a structured representation of voice, positioning, and creative standards that lets agents produce work a customer would still recognize as on-brand. The third is multi-agent orchestration, the connective tissue that lets several agents hand work to each other inside a campaign rather than each running in isolation. The fourth is talent, including engineers, prompt designers, and operators who can maintain the system once it is built. The marketing organizations that have invested in all four are the ones the report counts in the 8%.
The wider context sharpens the stakes. Nine in ten of the marketing officers surveyed say GenAI is already reshaping how consumers discover brands, particularly through AI-mediated search and recommendation surfaces that sit between the company and the customer. About half of the chief marketing officers now say the marketing organization, rather than the IT or data function, leads AI investment decisions inside the company, a sign of where the budget authority has migrated. The report also references a broader enterprise figure suggesting CEOs are increasingly positioning AI as a personal leadership priority, though the exact wording in the available excerpt is truncated and requires consulting the full source for precise attribution. The implication is the same: the marketing function is being asked to operate as an AI product organization whether or not it has built the underlying platform to do so.
The warning the report ends on is the part most likely to age well. If incumbent brands do not build the agentic operating layer themselves, the report argues, agentic-native challenger brands will build it first and use it to capture discovery, conversion, and loyalty in environments the incumbents cannot see. The source release points readers to a companion piece, "How CMOs are Moving Agentic Marketing from Illusion to Reality," for the same four moves in more operational detail. The framing is constructive, in the sense of naming a specific path, but it is also clear-eyed: most marketing organizations are not yet built to compete in the new discovery environment, and the marketing leaders who recognize that are the ones the survey identifies as the 8%.
The caveat that should sit alongside any of these numbers is that they are self-reported sentiment, not measured deployment or audited return on investment. The survey sponsor, sample size by geography, and fielding window are not fully visible in the press release excerpt, and the percentage contrast is shaped to land in a headline. A chief marketing officer reading the report as a roadmap is reasonable. Treating it as a market-share measurement is not.