58% expect agentic media buying within a year.

IAB Europe has just published its 2026 report on the impact of AI on digital advertising. The headline number: 58% of respondents expect agentic media buying to reach operational use within twelve months. In detail, 15 out of 50 see it becoming a main trading method in some markets, 14 expect day-to-day use without that level of scale. The report notes it itself: no scenario draws a majority, and 16% don't know.

One caveat before drawing anything from it. The sample is 50 responses, collected between 10 April and 29 June 2026: 35% agencies, 29% ad tech, 10% advertisers, 10% publishers, and 45% companies with fewer than 100 staff. IAB Europe states plainly that the findings should not be generalised to the European market. It is a climate indicator. We read it as such.

And read as such, it is telling, because the same survey gives the other half of the picture. Out of 47 respondents, 36 say they either have no agentic system in day-to-day use, or use one that people direct or plan with, levels 1 and 2 on the report's five-level autonomy scale. Agentic buying and selling is cited as an AI-supported workflow by only 11 respondents out of 29. And among the 21 who increased AI investment in 2026, only 7 have produced an advertising case study.

That is exactly what we see with the advertisers we meet. Everyone has understood that Google, Meta, TikTok and Amazon already operate a growing share of campaigns with their own agents. Almost no one has yet put an agent on their side of the table: one that reads the platforms at once, applies rules written by the advertiser, and reports in a common measure.

The rest of the report confirms the gap. 72% invest more than in 2025. 78% now have an AI governance owner. 62% ask their trade association for practical guidelines. Invest, organise, wait for a manual. Dimitris Beis, who authored the analysis, sums it up to ExchangeWire: "ambition remains ahead of practice".

Our reading: the next twelve months won't see agentic buying arrive all at once. They will see the gap widen between those who have framed it, a written plan, rules, a human who decides, and those who will undergo it through platform recommendations. The next two pieces look at who, within that expectation, is ahead, and at how the market measures what it already gets.