Agentic AI and the Next Evolution of AdTech

By Alistair Goodman, Cofounder, Emodo

AdTech has always had a distinctive strength: it adapts quickly.

Once new ideas, new acronyms, or new standards show promise, the ecosystem moves fast to adopt them. Sometimes that speed produces real innovation, other times, ambition arrives before execution. But historically, fast-following is how AdTech scaled, standardized, and kept pace with walled gardens.

The rise of header bidding or the industry’s rapid embrace of CTV are good examples. What looked like imitation was often survival — a way to keep innovation flowing on the open web.

But speed has a downside. Over time, fast-following can turn into accumulation.

Programmatic advertising is the clearest example. What began as a promise of automation, efficiency, and transparency gradually layered on intermediaries, abstractions, and operational complexity. SSPs, DSPs, data platforms, identity graphs, auctions within auctions — each layer solved a real problem in the moment. Collectively, though, they produced a system that became harder to understand, harder to optimize, and more expensive to operate.

It’s no accident that many of today’s corrective movements — supply path optimization, curated marketplaces, direct deals — are responses to that complexity. The industry is now working to simplify what it once worked hard to build.

That’s why Agentic AI announcements from PubMatic, Magnite, Index Exchange and others at CES represent the biggest structural shift in AdTech since the introduction of the iPhone. If the last era of AdTech was about building complexity to enable automation, the next era looks like automation designed to remove complexity altogether.

Agentic AI systems collapse planning, pricing, buying, optimization, and measurement  into a single learning loop. Instead of routing decisions through layers of tooling and teams, these systems evaluate inputs, act, measure outcomes, and improve continuously.

This isn’t about removing people from the process. It’s about removing the friction. And it brings the industry back to a simple question: what actually drives performance?

Search and Social have already begun to answer that question. Those channels don’t optimize media in isolation. They optimize systems where creative, audience, and delivery decisions are evaluated together against outcomes, in flight.

CTV is the last major performance channel still operating without that system-level intelligence.

For years, the focus has been on inventory scale, content quality, and reach. Those matter. But they aren’t what’s holding CTV back from performing like Search and Social. One of the biggest blind spots has been creative — not as an asset, but as a performance signal.

Most CTV campaigns still treat creative as fixed. When creative becomes a first-class optimization input — evaluated alongside audience and context — CTV starts to behave like performance media.

Agentic AI makes it possible to optimize creative, audiences, and inventory together — at scale, across the open web — with outcomes as the organizing principle. When that happens, CTV becomes easier to buy, easier to optimize, and easier to trust.

The real shift isn’t that machines replace humans. It’s that the way performance decisions are made changes. AdTech taught the industry how to automate buying and selling. The next phase is about applying that same discipline to creativity and outcomes, and building performance systems that finally live up to the promise of TV.