The Year of Agentic AI: How’s That Going?

We’re six months into what many predicted would be “the year of agentic AI.”

Looking at the pace of innovation, it’s hard to argue that AI hasn’t become the defining conversation of 2026. Every week seems to bring another product launch, another capability and another vision for how autonomous media buying might become.

And honestly, that’s exciting.

Technology is moving at an incredible pace, and much of what’s being built today would have felt impossible just a few years ago. AI is already helping buyers work faster, uncover insights more quickly and remove many of the repetitive tasks that have traditionally slowed campaigns down.

But beyond the announcements and conference stages, there’s a more interesting question.

How much has actually changed for the people buying media every day?

Walk through any adtech event today, and it’s clear the industry is investing heavily in agentic AI. DSPs, SSPs, measurement partners and data providers all have a vision for how AI can improve planning, activation, optimisation and reporting. That’s a good thing. Innovation has never been in short supply in this industry.

Speak to buyers, however, and the conversation is often a little more - let's say - measured.

Not because they doubt the technology, but because they’re still working within the same fragmented ecosystem they’ve always known. And it’s proving hard to escape. 

Planning happens in one place. Activation happens somewhere else. Measurement lives in another platform entirely. AI is making individual parts of the process smarter, but the overall experience can still feel disconnected.

That’s where I think the conversation becomes really interesting.

The biggest opportunity isn’t simply making each tool more intelligent. It’s making the entire media buying process work together more effectively.

More signals. Same confidence gap.

Many buyers still find themselves asking the same question at the end of a campaign: Did this actually achieve what I set out to do?

That’s not because the industry lacks data and insight. Buyers have never had more of it. It’s because the data rarely exists as one connected story.

Signals sit in one system. Activation happens in another. Measurement often uses a different definition of success than the one the campaign was originally built around. Every platform is optimising its own part of the process, but very few are helping buyers understand how those parts come together to deliver a meaningful outcome.

That’s the confidence gap. Which is why I think the trend that matters most this year isn’t AI on its own. It’s connected decision-making.

When trusted signals, quality inventory, activation and measurement are brought together around a shared objective, AI becomes significantly more valuable. It isn’t simply making decisions faster; it’s helping buyers make better decisions with greater confidence.

As such, connected decision-making no longer feels like delving into an overwhelming pit of complexity, and instead feels more like a ray of sunshine; a single, connected process designed around the outcomes that matter to you.

Where we think this goes next

As AI becomes embedded across every stage of media buying, I think the conversation naturally shifts from what AI can do to how buyers can trust what it’s doing.

• Can the plan explain itself? Buyers need to understand why a recommendation has been made, not simply accept it because an algorithm suggested it. Explainability is quickly becoming the difference between a useful assistant and a black box.

• Can performance be measured against the outcome that actually mattered? Not every campaign exists to maximise clicks or viewability. Metrics like attention or relevance only become valuable when they’re connected to the objective the buyer set at the very beginning.

• Can the results be validated once the campaign is over, in a way that gives buyers confidence beyond the report that generated them?

None of those questions requires another platform.

They require the platforms, signals and technologies that already exist to work together far more effectively than they do today, creating a connected feedback loop where every campaign helps improve the next.

That’s the direction we’ve built @curate around.

Rather than adding more complexity to an already crowded ecosystem, we’ve focused on connecting the infrastructure. Buyers define the outcome they’re trying to achieve. Wanda (our AI planning agent) builds the strategy around that objective. Aardvark (our connected media buying system) activates and continuously optimises campaigns across display, video, audio and CTV. And, then our reporting closes the loop - explaining not just what happened, but why it happened and what can be learned from it.

No new platform to learn (we do all the heavy lifting). No fragmented workflows to stitch together. No additional operational complexity layered onto already busy teams.

Just a connected process built around delivering the outcome that mattered from the start.

The first half of 2026 has understandably been dominated by conversations about how autonomous AI agents can become. However, I suspect the more important conversation for the second half of the year is how connected the entire media buying process can be.

Should be, even.

Because AI, on its own, won’t define the future of media buying.

Confident and connected decision-making will.