CNN Builds Agent Trading Infrastructure for 2027 Launch
CNN is developing its own agentic infrastructure to begin transacting media through AI agents by the first quarter of 2027. The news organization plans to finish scoping agentic protocols by the end of Q2 2026, then test how properties are interpreted by large language models in Q3, according to Faisal Karmali, VP of digital business operations at CNN International Commercial.
The timeline remains flexible as agentic protocols continue to develop, Karmali said. CNN will pair in-house development with third-party tech and trading vendors, including demand-side platforms, while working with the IAB Tech Lab to ensure protocols align across the industry.
Learning Without Revenue
CNN's approach differs from traditional programmatic ad tech adoption. Rather than waiting for buyer-side adoption before investing, the publisher plans to implement agent protocols now-even without immediate revenue flowing through them.
"This is a learning protocol," Karmali said. "If we implement it now and there's no revenue being traded, when it does start trading, it'll be super efficient. It'll kind of leapfrog that learning step."
The focus centers on agent-to-agent communication: how buy-side and sell-side agents request information, negotiate pricing and usage terms, and delegate tasks. Consistency across these interactions is critical.
Early Buyer-Side Results
Tests on the buy side show promise. Media agency Butler/Till recently concluded an experimental programmatic media-buying agent campaign that cut intermediary fees by over 80% and reduced CPMs while maintaining industry fraud and inventory standards.
News Corp is also developing internal agentic infrastructure for its titles. However, most sell-side examples focus on campaign setup rather than actual trading. Early cases are likely to involve guaranteed, pre-arranged deals across multiple publishers rather than open real-time auctions.
Multiple Agents, One Market
Publishers can run their own sales agents alongside external ones, creating what some call "agentic abundance"-multiple agents operating in parallel. Publisher-specific agents sell only their inventory, while broader platform agents aggregate supply across multiple publishers.
This model gives publishers control and visibility into who's buying, allowing them to identify high-value buyers for direct outreach and partnership opportunities.
Starting With Performance
Agent-based trading will likely prove most effective initially for lower-funnel, performance-driven activity like conversions and direct response. Once payment and commerce capabilities are integrated, the technology can move upward to handle branding and discovery campaigns.
CNN views the shift as an opportunity to establish new trading standards from the ground up. "We've got an opportunity here to kind of start from scratch, start a new way of trading, which we don't get often," Karmali said.
The publisher remains open to conversations with buy-side partners about preferred protocol approaches. Aligning on consistent context protocols across both sides is essential for efficient agent trading.
Learn more about AI Agents & Automation and how Generative AI and LLM technologies are reshaping industry operations.
Your membership also unlocks:
