CNN International Develops Agentic Ad Market Infrastructure

CNN International is moving forward with its plan to sell ads through AI agents, with a launch expected in the first quarter of 2027. The network's vice president of digital operations, Faisal Karmali, expressed hope that the first transactions would take place by then. According to Digiday, the company has been working on this project for seven months.

The network is building its own agentic infrastructure in-house, while also working with third-party technology vendors. This infrastructure will be used to sell ads through AI agents, which will require the development of industry frameworks and protocols. CNN International's plan is to start with performance campaigns before expanding to other areas.

Karmali provided an update on the project at Digiday's Media Strategies event in London, stating that the team is focused on building a functional system before seeking deals. He noted that there is still much work to be done, particularly in terms of teaching the AI agent the company's business rules and ensuring that its inventory is readable by large language models.

One of the challenges facing CNN International is the lack of standardization in the industry, with competing standards and no agreed-upon way to prevent agents from gaming the system. Karmali emphasized the need for open communication between buyers and sellers to address these issues. Digiday reported that the company is working to establish a common language for trading with agencies and clients.

The potential benefits of this system for publishers include more control over how their inventory is sold and what it sells for. However, the industry has yet to agree on what constitutes an accurate description of inventory, which could impact the effectiveness of the system. CNN International believes that its high-quality content and clean metadata layer will give it an advantage in this area.

The company's goal is to create a system that can efficiently trade ads between buyers and sellers, without the need for human intervention. This would require the development of a common language and standards for the industry, as well as a way to prevent agents from misbehaving. While the potential rewards are significant, the challenges are substantial, and it remains to be seen whether CNN International can successfully navigate them.