Van Zwol outlined Reddit’s emerging approach to agentic traffic. Rather than blocking agentic traffic outright, the platform is focused on identification and containment, aimed at ensuring agents make themselves known and operate within defined limits. “We do want to make sure that when there are agents, we limit what they can do, or at least that they make themselves known,” he said.
The question of trust also surfaced around content quality and LLM scraping. Tripadvisor, which holds more than a billion human-generated reviews, has found itself in an unexpectedly strong position as AI systems increasingly favor sourced, specific, and recent content, said Allison Lambroza, senior director and head of media and brand partnerships at Tripadvisor.
“We are the number one most cited travel source in over 70% of queries,” Lambroza said. “We still see over 40% of people citing misinformation as a barrier to entry when it comes to travel. People still want to do their research.” For Tripadvisor, she said, the flywheel is the same one that built the company: trusted reviews from real people feed LLMs that reward specificity.
CTV’s different equation
Connected television is not immune from bots, but structurally harder to game, according to Lauren Benedict, vp of global ad sales and partnerships at Roku.
“The CTV space, it’s just a little bit more expensive for bots to live in because tracking video is challenging,” Benedict said. Roku’s 100 million authenticated, logged-in households give it a foundation that open web publishers lack, but Benedict was clear that the platform isn’t complacent. Roku has partnered with Human and built watermarking capabilities designed to validate that a real viewer is on the other side of the content.
Interoperability is central to how Roku is building for what comes next, Benedict said—constructing data and tech infrastructure that can support agent-to-agent buying as it matures, while preserving the human accountability layer that advertisers still require. “The human touch is still important to us,” she said. “But the fair amount of build that we’re doing to integrate is really meaningful.”
Who’s accountable when agents transact?
Citing recent data from the Interactive Advertising Bureau, ADWEEK’s Barnett noted that up to two-thirds of U.S. media buyers are already exploring investment in autonomous campaign execution. This raises the question of: when something goes wrong in a fully automated transaction, who owns it?
Human’s Solomon called out Google’s framework as a good example of how to approach the issue: agents must be authentic, authorized, and accountable. “The accountability starts with the person, the human being, who set it on its journey in the first place,” he said. “Whether it’s within the advertising mindset or the broader economy of the internet, that accountability is going to further sit with the human being who set the plumbing in motion.”
Reddit’s Van Zwol agreed that the feedback loop itself will become a form of accountability. “Rather than running a campaign and waiting three weeks to get results, that feedback loop becomes much quicker,” he said. KPIs, Tripadvisor’s Lambroza added, will shift accordingly—away from page views and toward intent and qualification.
