Agentic Buying: Automated Media Buying Arrives in 2026

What agentic buying is and why everyone is talking about it now

Agentic buying it means media buying handled by autonomous AI agents: software that plans, buys, and optimizes ad space without waiting for human input on every single decision. This isn't just an extra feature inside a platform; it's a structural shift in how budgets move.

According to an IAB Europe study reported by Digiday , 58% of ad execs expect this model to hit operational scale within a year. The industry comparison everyone's making is to mobile: for years people said it would become the main channel, until one day it actually was and anyone who wasn't ready was left playing catch-up.

The source numbers and what they actually mean

The main takeaway is clear: more than half of industry executives believe that agentic buying is not a side experiment but an upcoming evolution. IAB Europe is the European trade association for the digital advertising industry, so the people answering these surveys are the ones calling the shots on budgets and strategies, not academic researchers.

What does that 58% actually mean in practice?

• Platforms are already building infrastructure for autonomous agents — they are not waiting for the market to be ready.
• Those who buy media on behalf of client companies — agencies, in-house media buyers — will see their role change sooner than they think.
• Those who sell spaces — publishers, agencies — will need to adapt their APIs and formats to talk to software, not just people.

Who this week's changes matter for — and who they don't

If you're running campaigns with small budgets and are still manually optimizing ads on Meta or Google, agentic buying won't affect you tomorrow morning. However, the platforms you already use are integrating increasingly advanced automations: what you currently call "smart bidding" or "Advantage+" is the precursor to this model. Understanding the direction helps you better interpret the choices Meta and Google are making on their platforms — as highlighted in the analysis on how Meta is shaping its 2026 holiday campaigns with AI .

If instead you manage significant budgets across multiple channels, have a media buying team, or work with an agency, the issue is already operational:

• Are your campaigns structured so that an agent can read and optimize them? Clean data, clear naming conventions, measurable goals.
• Does your team know where human control needs to stay and where they can delegate?
• Are your ad tech vendors heading in this direction?

The real risk: delegating without understanding what you're handing over

The most common mistake with any new advertising automation is activating it without defining the boundaries. With agentic buying, the risk is amplified: an autonomous agent can optimize towards metrics that don't align with your real goals, distribute budget across inventory (available ad spaces) that you haven't validated, or react to short-term signals that harm brand positioning.

This is not a theoretical problem. It is the same issue already showing up today with Google's Performance Max campaigns — where the algorithm decides where to go and the advertiser's control is reduced. The issue of Customer Lifecycle Goals in Google Ads it is a concrete example of how platforms are already shifting control toward the machine, and of when it makes sense to follow them and when it does not.

The same logic applies to new formats: ChatGPT click-to-chat ads operate on a completely different logic compared to traditional site traffic, and they require redefined goals even before being activated.

Three things to do before the shift is fully in place

You don't need to wait for agentic buying to become the standard to start preparing. Three concrete actions, in the order that makes sense:

• Clean up campaign data. Autonomous agents work on signals: if your data is fragmented, poorly labeled, or lacks correctly tracked conversions, automation optimizes into a void. Start here, regardless of everything else.
• Set your goals with measurable targets. "Increase brand awareness" is not a goal an agent can optimize for. Cost per lead, average order value, repurchase rate: these are. The topic of Hidden costs of AI in marketing shows well how automation without clear goals erodes efficiency instead of increasing it.
• Map out where you want to keep human control. Creativity, brand positioning, sensitive context exclusion choices: these are areas where human supervision is not a cost, it's a protection.

For those who want to dive deeper into the big picture of how this is evolving digital advertising and paid media , the thematic cluster gathers the most relevant updates for those managing campaigns in Italy.