58% of ad execs expect agentic buying to hit scale within a year, according to IAB Europe research
Ad execs have kept their distance from agent-to-agent ad trading, wary of the technical cost and definitional questions still hanging over it. New research says that wariness is starting to shift.
Fifty ad execs across 44 European markets were asked about it this spring and summer, and 58% of them expect agentic ad buying to hit operational use or scale within a year. What that phrase actually means depends on who was asked, because the businesses in that sample aren’t starting from the same place. Unsurprisingly, the bigger organizations are well ahead. Most of them (86%) who are at companies with 501 or more staff already run their most advanced agentic systems at a level where people and agents plan, delegate and execute work together. Drop down to companies with 500 staff or fewer and that number falls to less than half, 48%.
Expectations of agentic trading are just as divisive, if not more so.
Most execs expect further development but nobody agrees on how quickly AI agents will actually become part of everyday ad trading. Three in ten (30%) of those surveyed expect AI agents to become a main way of buying and selling ads in some markets, while more than a quarter (28%) expect regular use without that level of scale.
Regardless of where they sit on that sliding scale, the direction of travel is clear for ad execs. Agentic advertising is coming for everyone, just not at the same pace.
Forty seven of the 50 respondents said their organizations already use AI internally, and slightly less (43) use it specifically for marketing.
That broad outline matches the anecdotal evidence that’s built up throughout the year. Holdcos are arguably further down the road than most. WPP has built and is currently testing its own buyer agent. Omnicom is doing the same. On the other side of the table, publishers are moving too: CNN and News U.K. are building sell-side agents to handle those requests. Advertisers, for their part, aren’t ruling out building their own either but the ones like Georgia-Pacific want more proof the industry is ready before they commit.
None of them are losing sleep over whether AI gets used. What worries them is how well it works and who’s actually steering it.
Thirty six of 47 respondents said as much, answering that their organization either has no agentic system running day to day or one that people still direct or plan alongside. And it’s hard to blame them. Standards haven’t caught up yet, the logistics remain messy and handing a machine that much autonomy with so little oversight still makes plenty of people nervous. Of those concerns, security is top, according to the data. Privacy is second.
Which is why, for now at least, AI agents get put to work on the more monotonous parts of the job. For instance, 22 of the 29 respondents selected reporting, analysis and dashboards as the most common AI-supported workflow in their business. Programmatic optimization came next, with 59% of the vote.
Tasks like those are graded on output, not attribution, so it tracks that efficiency topped the list of how those same respondents said their organizations judge AI. Seven in 10 (68%) of those who responded to the whole survey said as much.
That said, their stance will shift as agents take on more tasks. The economics demand it. Efficiency gains save money and that money gets reinvested into making the agents faster, better and cheaper. Most respondents are already spending accordingly. They’re putting more into AI in 2026 than they did in 2025, and 59% are increasing investment in AI marketing technologies specifically.
Ominicon’s cfo Phil Angelastro flicked at that very point earlier this month at the Goldman Sachs’ Communacopia and Technology Conference. He said: “There is going to be likely more investment from advertisers because one of the meaningful changes that’s been happening in the business — and that will continue to happen with AI — is all of what we do and results are going to be much more clearly and easily measured. Some of them, it’s going to be a challenge to measure, but ultimately measurement is going to improve and the more it improves the better it’s going to be for us because we’re going to be able to prove the return on investment that we deliver to clients.”
Here’s the thing with all the talk about agentic buying. Whether or not it works at scale is secondary to if it is able to move the needle for CMOs. Not everyone thinks it will. Skeptics argue the whole conversation automates a sliver of the process while leaving the real bottlenecks untouched.
“I always ask what value is agentic buying bringing?,” said Ian Maxwell, CEO at ad tech business Converge Digital. “It automates a few small parts of a huge process yet still needs human monitoring and oversight. It won’t have a substantial uptick in outcomes and performance if we’re just training it on past buying data so automating what’s happened previously. And even when the campaign takes seconds to create from start to finish….. when the creative is 4 days late, what exactly is different to right now?
More in Media Buying
Talent firm 24 Seven consolidates the new business efforts of its three concentric agencies
SktechDeck, Markacy and Futureman are now being pitched by 24 Seven’s new business team as any combination of the three shops that a client might want to tap
Ad Tech Briefing: Google’s ad tech remedies arrive just as the platform battle moves to AI
Regulators are prising open Google’s programmatic machinery just as AI, commerce and ad tech players race to control more of the advertising stack.
Can adland realistically reduce its reliance on Google’s vertically integrated ad stack?
The court has ordered Google to open up its ad tech stack, but can interoperability alone can loosen its grip on the market.