IAB Europe has found that expectations for agentic media buying are advancing faster than the operating systems around them. Advertising executives increasingly expect AI agents to move from experiments into daily buying workflows, even as security, privacy, training, and accountability remain unresolved.
This is not simply another adoption story. It is evidence that media organizations may soon delegate more campaign decisions before they have agreed on how autonomy should be governed or how its value should be measured. Agentic buying is becoming operational before it becomes institutionally settled.
58% of respondents expect agentic ad buying to reach operational use or scale within the next year.
Table of contents
Jump to each section:
- Agentic buying is moving from pilot to production
- The readiness gap is organizational, not just technical
- Efficiency is a weak finish line
- What marketers should know about agentic media buying
Agentic buying is moving from pilot to production
The research places agentic advertising between two stages. AI is already common in marketing operations, but systems that can plan, delegate, and execute alongside people are still unevenly distributed. The expected direction is clear, while the degree of autonomy is not.
That distinction matters because the category covers very different realities. An agent that assembles a dashboard is not carrying the same commercial risk as one that changes budget allocation or transacts with a sell-side system. Treating both as evidence of agentic maturity makes adoption look more coherent than it is.
Live market activity reinforces the shift. WPP and Omnicom are testing buyer agents, while CNN and News UK are developing sell-side agents. These examples move the discussion beyond chat interfaces and into the transaction layer, where software may increasingly act on a marketer’s intent rather than merely interpret it.
The deeper change is not faster button-clicking. It is the transfer of decision rights from interfaces to systems.
The readiness gap is organizational, not just technical
Larger organizations appear much further along in combining human and agent-led work. That advantage is likely to reflect more than access to stronger models. Large operators can spread the cost of data integration, security reviews, workflow design, and specialist talent across more campaigns and clients.
86% versus 48% is the gap between larger and smaller organizations reporting their most advanced production agentic systems at a collaborative level where people and agents plan, delegate, and execute work together.
Scale may improve the machine before it improves the organization. A company can deploy more agents and still lack clear ownership when a campaign violates a constraint, uses the wrong signal, or optimizes toward a misleading proxy.
The common assumption is that smaller teams will benefit first because automation compensates for limited headcount. The contrasting reality is that larger businesses currently have more capacity to build the controls that make autonomy usable. The strategic implication is that agentic buying could widen operating gaps unless vendors make governance, auditability, and integration accessible to teams without extensive technical resources.
This also changes the vendor question. Marketers should look beyond what an agent can do and ask what the surrounding organization must supply: clean data, permissions, exception handling, approval logic, and a reliable record of decisions.
78% and 48% of respondents respectively report a formal owner for AI governance and advertising-specific AI guidelines, revealing a gap between assigning responsibility and defining rules for marketing use.
Efficiency is a weak finish line
The current use cases are revealing. Reporting, analysis, dashboards, and programmatic optimization are natural places to start because the work is repeatable and outputs can be reviewed. They also sit close enough to execution to create visible time savings without always requiring an agent to own the full commercial decision.
Yet efficiency is an incomplete measure of whether the system improves marketing. A campaign can be assembled faster while the creative still arrives late. An agent can rebalance spend more often while continuing to learn from historical buying patterns that reproduce the same blind spots.
69% of respondents selected operational efficiency as a measure of AI effectiveness, making it the most commonly cited yardstick in the study.
When efficiency is the primary scorecard, automation can look successful while marketing effectiveness remains unresolved.
The more demanding standard is whether agentic buying improves decisions that matter to a CMO: better allocation across channels, clearer links to business outcomes, more reliable compliance, and faster learning that changes future strategy. Those outcomes require measurement design and organizational judgment, not just technical speed.
Security and privacy concerns therefore belong inside the performance conversation. They are not side constraints to be handled after deployment. If a system cannot explain what it accessed, changed, or optimized, the apparent gain in execution speed may create a larger cost in review, reconciliation, or trust.
What marketers should know about agentic media buying
The near-term opportunity is real, but the most useful preparation is to define the operating model before expanding autonomy.
Separate assistance from authority. Map which systems generate recommendations, which can execute changes, and where human approval remains mandatory. The word “agent” is too broad to serve as a control policy.
Measure decisions, not only task completion. Time saved on reporting is useful, but marketers also need to know whether an automated choice improved campaign quality, business outcomes, or the speed of learning.
Design for exceptions. Media systems become strategically important when conditions depart from the expected pattern. Teams need ownership, audit trails, and override paths for the moments an agent cannot resolve cleanly.
Treat interoperability as a commercial issue. Buyer agents and seller agents will need shared ways to exchange objectives, constraints, and transaction details. Without that coordination, automation can add another fragmented layer to the adtech stack.
The competitive advantage may not be the agent. It may be the operating discipline around it.
As agentic buying matures, many platforms will offer similar claims about speed and optimization. Marketing organizations will differentiate through the quality of the goals they encode, the limits they set, and the evidence they demand before giving a system more authority.
That shifts the role of media leadership. The work moves upward from managing every action toward designing the environment in which actions can be delegated responsibly. Agentic advertising will scale through technology, but it will create durable value only when governance, measurement, and human judgment scale with it.