Agents of Change at Cannes Lions: Where Do You Trust an Agent Today?

The Agents of Change interview series, recorded on the Croisette at Cannes Lions 2026. Produced in partnership with AgenticAdvertising.org. Sponsored by Agnitio.

If you spent any time at Cannes Lions this year, you heard the same thing we did: the industry is done debating whether agentic advertising is coming. It's here. We see it in production campaigns, in creative pipelines, in the operational plumbing of major media companies. The question that kept surfacing in every panel, mixer, and Croisette conversation was sharper: agents are acting on our behalf right now, so where, exactly, does human judgment live?

That question is the reason this series exists.

Why we made this

AAO's thesis is simple: human judgment isn't a bottleneck slowing agentic AI down, rather, it's the architecture that makes scale trustworthy. Speed without judgment is just fast mistakes. And this year, Cannes gathered the people proving it in production: the leaders whose agents are already planning campaigns, transacting inventory, and generating creative at scale.

So we put them on the record. Together with the Agents of Change podcast, sponsored by Agnitio, we sat down at the iHeartMedia Cafe with ten leaders spanning every seat in the ecosystem — buy side, sell side, creative platforms, governance and security, publishers, and the standards body itself: Brian O'Kelley (Scope3), Priti Ohri (Advertible), Ben Williams (Kiln), Gabriel DeWitt (Yahoo), Matic Tribuson (Celtra), Mike Chevallier (iHeartMedia), Pamela Slea (Boltive), Chinmayi Bettadapur (Salesforce), Sarah Acker (HUMAN), and AAO President Benjamin Masse.

The conversations trace a few themes, and we'll be releasing them as short reels over the coming weeks: where these leaders trust agents in their work today, where human judgment remains non-negotiable, what advertising's jobs look like a decade out, and more. The first reel is live now and it starts where trust has to start: with what's actually running in production.

Episode one: trust is already here, and it's specific

Watch it here: three and a half minutes, eight leaders, one question: where do you trust an agent in your work today?

What struck us most is how concrete every answer was. Nobody spoke in hypotheticals.

Brian O'Kelley gave the most personal version. After three decades of writing programmatic's code with his own hands, he now runs 15 to 20 projects through agents at once. "I have a QA agent that QAs my coding agent. I have a design agent, I have a docs agent... I don't tell it how to code it, I tell it the objective I want, and then I come back an hour later."

From there, the reel moves across the ecosystem:

Salesforce's Chinmayi Bettadapur described agents fast-tracking a long-wished-for product roadmap by letting product teams prototype hands-on. She describes this as "music to my ears."

Mike Chevallier walked through iHeartMedia's sales modernization in progress: agents planning broadcast radio through their Audiograph identity product, an MCP already in front of digital inventory for agentic testing in podcast and streaming, and the same treatment coming to their radio scheduling system to open broadcast inventory to agentic transactions.

Ben Williams of Kiln drew his trust line at the platform level: agents running autonomously across integrations he already trusts: Workspace, Slack, and the stacks of providers like Anthropic.

Priti Ohri put it simply: at Advertible, agents amplify the voice of the company itself through marketing, in coding, and live in production inside the product.

Gabriel DeWitt brought the receipts from the sell side: production campaigns at Yahoo with a major holding company, buyer and seller agents transacting through the Ad Context Protocol. He calls it "agentic direct" — the agentification of the insertion-order workflow, publisher-to-buyer, not programmatic wearing an AI costume.

Pamela Slea showed what scale actually means: Boltive's agents checking consent compliance across hundreds of thousands of sites, jurisdiction by jurisdiction. "Without agents, we wouldn't be able to do that."

And Sarah Acker closed with a brand story that captures the whole moment: HUMAN rebuilt its own brand identity using AI-generated imagery — humans, good bots, and bad bots coexisting — because in the agentic era, "bot" no longer means "bad." A brand called HUMAN, reimagined with machines. That tension is exactly what this series is about.

What's coming next

Trust is only the first of our themes. In the reels ahead, the same leaders take on the harder half of the equation: where human judgment remains non-negotiable — from iHeart's "fit, money, and law" test to Yahoo running one campaign with zero humans purely to prove it could, to Benjamin Masse's case for the standard AAO builds toward: Embedded Human Judgment, approval built into the process before an agent acts, logged and traceable, not discovered after the fact.

And then we go to 2036, where ten predictions about disappearing jobs converged on the same list — trafficking, ad ops, report-pulling, form-filling — and the new roles got names like agent orchestrator, VP of Agentic, curator, and O'Kelley's favorite: the dreamer.

Ten guests, ten vantage points, one conclusion: the technology is ready to scale faster than any human team, and whether it earns the industry's trust depends entirely on whether human judgment is built into how it operates — not layered on top afterward.

Watch episode one Agents of Change Podcast | Where Do You Trust an Agent? | AAO at Cannes Lions 2026

Go deeper: Read AAO's whitepaper, Building the Future of Marketing, and explore the Ad Context Protocol at agenticadvertising.org.

Join us: AAO membership is open to brands, agencies, publishers, and technology companies who want to shape the future of AI and Advertising. Become a member today.

Agents of Change was produced in partnership with AgenticAdvertising.org and sponsored by Agnitio.