Agentic AI Is Coming to Media Buying — But Accountability Has to Come First

Agentic AI is quickly moving from industry concept to advertising product roadmap.

Recent trade coverage has pointed to AI agents being developed for media planning, campaign modification, buying and optimisation. The Trade Desk has introduced AI-powered agents for media planning and buying, with Stagwell piloting the capability. Netflix is also reportedly introducing AI tools and agents to support advertising growth, including planning, optimisation and creative adaptation.

At the same time, Digiday has covered the debate around agentic media buying: whether it could simplify programmatic’s supply chain problems or add a new layer of opacity to an already complex ecosystem.

The opportunity is clear. But so is the risk.

What happened

Programmatic advertising has always been shaped by automation. Bidding, targeting, optimisation, pacing and reporting have all become increasingly automated over time.

Agentic AI takes that one step further.

Instead of simply optimising within rules set by humans, AI agents may begin making more autonomous decisions across planning, buying, campaign adjustments, reporting and creative variation.

That could make media buying faster and more efficient. But the commercial question is not just whether an agent can optimise a campaign.

The better question is: who is accountable when the agent optimises toward the wrong outcome?

Why it matters for AUNZ

Australia and New Zealand have sophisticated media and programmatic markets, but they are also relatively concentrated.

A small number of major media agencies, publishers, platforms, broadcasters, retail media networks and technology partners shape much of the market. That makes trust, transparency and accountability especially important.

If agentic AI becomes part of local media buying workflows, advertisers will need confidence around where budget is being spent, what inventory is being selected, which data signals are being used, how decisions are optimised, whether brand safety is protected, how performance is measured, and who is responsible if something goes wrong.

This is particularly relevant in areas like CTV, retail media, programmatic display, digital audio and platform advertising, where measurement and supply-path questions are already complex.

My view

My view is that agentic AI will become part of media buying, but it should be treated as a workflow layer before it is treated as a decision-maker.

The early value is likely to come from helping teams with planning inputs, scenario modelling, reporting, optimisation recommendations, budget pacing, creative variation and campaign analysis.

That is different from allowing an agent to independently decide where money goes without human oversight.

The history of programmatic shows why this matters. Automation can create scale and efficiency, but it can also create distance between advertisers and the actual media value being bought.

If agentic systems add another layer between the advertiser and the impression, then the industry risks repeating old mistakes under a new AI label.

The goal should not be automation for its own sake. The goal should be better decisions, clearer accountability and more useful outcomes.

What I’d be watching

First, I’d watch how platforms explain the role of human approval. Are agents recommending actions, or taking them automatically?

Second, I’d watch how agencies describe accountability to clients. If an AI agent changes campaign settings, buying paths or budget allocation, who signs off?

Third, I’d watch how transparent these tools are about inventory, data and optimisation logic. If the answer is simply “the AI decided,” that will not be enough for serious advertisers.

Fourth, I’d watch whether agentic AI improves supply-path clarity or makes it worse. Programmatic has spent years trying to reduce unnecessary complexity. AI should not reintroduce opacity.

Finally, I’d watch how quickly this moves into AUNZ media plans. The local market may not lead every platform rollout, but Australian buyers are experienced enough to ask sharp questions once these tools arrive.

Question for the market

Agentic AI may make media buying faster.

But will it make media buying better?

That depends on whether the industry can answer a more important question first:

Who is accountable for the decisions the agent makes?

Sources

• Digiday: The case for and against agentic media buying — digiday.com/media/the-case-for-and-against-agentic-media-buying/

• Digiday: Accountability in adland’s agentic era — digiday.com/media-buying/ad-tech-briefing-accountability-in-adlands-agentic-era/

• Adweek: The Trade Desk AI agents coverage — www.adweek.com/programmatic/the-trade-desk-growth-slows-to-lowest-rate-since-covid-with-publicis-negotiations-ongoing/

• Adweek: Netflix AI agents coverage — www.adweek.com/media/ai-agents-are-coming-to-netflix-to-grow-its-3-billion-ad-business/

• IAB Australia: A Guide to Prompting with LLMs 2026 Edition — www.iabaustralia.com.au/a-guide-to-prompting-with-llms-2026-edition/