Seventy percent of respondents already run custom AI agents handling real marketing tasks in production, according to a recent report by Kana.
Widespread adoption — and the new question of scale
Kana’s survey finds adoption is no longer the central issue. Only 3% of respondents report no marketing AI use at all, and 70% say they have custom AI agents in production performing real marketing tasks. That widespread deployment, the report argues, shifts the executive debate from “should we adopt?” to “how do we scale and govern what is already running?”
The practical consequence is immediate: organizations that have already embedded agents must now address operational and organizational challenges that simply did not exist at the earlier, exploratory stage of AI use.
Ownership fights: Chief AI Officer, marketing, or shared control?
Across the full sample, 40% of respondents say the Chief AI Officer should own agentic marketing strategy and execution. The split differs by respondent type: “AI leaders” are more decisive, with 52% pointing to the Chief AI Officer (CAIO). Marketing executives, by contrast, lean toward their own function and toward shared ownership.
Kana warns of a practical risk: until enterprises resolve who is accountable, “agentic marketing initiatives risk stalling in the gap between teams that each assume someone else owns the outcome.” The result can be duplicated effort, misaligned priorities, or stalled projects even where the technical capability already exists.
Readiness metrics — and a striking gap with reality
Leaders in the report are relatively confident on two fronts: 76% say their governance model is ready for supervised AI decisions, and 86% rate their data infrastructure as ready. Yet those same leaders list data governance readiness and data quality/hygiene as their second and third biggest obstacles to agentic marketing progress.
That contradiction — high self-assessed readiness alongside naming data governance and quality as leading barriers — is a core tension in Kana’s findings. In short: infrastructures and governance frameworks may look prepared on paper even as underlying data problems threaten to blunt agent performance and produce unreliable outcomes in production.
Expectation: agents will manage a substantial share of decisions
Respondents are betting on a rapid shift in who decides routine marketing matters. According to the report, 82% expect AI agents to handle at least a third of routine marketing decisions within two years; 46% expect a majority of those decisions to be handled by AI agents.
Those projections are driving investment and planning even where operational gaps remain. Kana notes the primary investment driver is competitive anxiety: 69% of respondents say concern about falling behind peers outweighs every other risk, including security and data privacy. In other words, fear of being outpaced by rivals is trumping traditional risk calculations.
What this means for security teams, marketing executives, and chief AI officers
- Security teams: The report’s sidebar — “ON DEMAND: Security teams have never had more visibility” — suggests security groups are being given more contextual tools. As organizations shift from detection to decision, security teams will need to evaluate agent actions and data flows to reduce reliance on manual triage and enable faster action.
- Marketing executives and managers: Marketers are the cohort most likely to favor retaining control or sharing ownership. They will have to balance protecting brand and customer relationships against pressure to cede routine choices to agents that promise speed and scale.
- Chief AI officers and enterprise leadership: The CAIO role is the plurality choice for ownership of agentic marketing. If leaders consolidate accountability there, CAIOs will face the twin tasks of proving governance models work in practice and closing data-quality gaps that their peers have identified as major obstacles.
Kana’s data draws a clear picture: enterprises are already running agentic marketing at scale, expect agents to take on far more decision-making soon, and are moving forward primarily because they fear competitive disadvantage. Yet confidence in governance and infrastructure coexists with recognition that data governance and hygiene are significant barriers. Until organizations resolve ownership and clean their data, the transition from pilot to durable production risks being driven more by anxiety to keep pace than by confidence in safe, reliable outcomes.
