The Future of Ad Tech: CNN International’s Bold Bet on Agentic Workflow

By Seb Joseph
October 22, 2026

As the media industry grapples with the seismic shifts brought on by the Skydance deal and a rapidly evolving digital landscape, CNN International is placing a high-stakes bet on the future of advertising: AI-driven agentic workflows. Despite the corporate restructuring and uncertainty surrounding parent company Warner Bros. Discovery, CNN International’s roadmap for automating its advertising sales remains steadfast.

Seven months after announcing its pioneering move into AI-based media trading, the network is officially on track to launch its in-house agentic infrastructure in the first quarter of 2027. For an industry accustomed to manual, labor-intensive Request for Proposal (RFP) processes, this pivot represents a potential paradigm shift in how premium inventory is bought, sold, and valued.

The Strategic Timeline: A Q1 2027 Launch

Faisal Karmali, the vice president of digital operations for CNN International Commercial and Warner Bros. Discovery, remains optimistic about the project’s trajectory. "I gave a Q1 timeline to have the first dollars flowing, which I would still be hopeful for," Karmali noted during a recent industry appearance at Digiday’s Media Strategies event in London.

The development process has been methodical. Throughout 2026, the network has been quietly building its internal agentic architecture, integrating it with third-party technologies and demand-side platforms (DSPs). The internal roadmap was structured in phases:

• Q2 2026: Scoping and defining technical protocols.
• Q3 2026: Testing how LLMs (Large Language Models) interpret various CNN properties and content types.
• Q4 2026: Analyzing buyer behavior, budget allocation patterns, and fine-tuning the agent’s decision-making logic.

The strategy is to begin with performance-based campaigns before scaling into broader, more complex brand-building inventory. While the "switch-on" moment for live trading has not yet arrived, the infrastructure is reaching a level of maturity that allows the network to simulate how a "buy-side agent" interacts with their inventory.

The Complexity of Agentic Infrastructure

Building a functional agentic system is not merely a software challenge; it is a fundamental restructuring of how media business rules are coded and executed. CNN International is currently navigating a "Jenga-like" landscape of technical and operational hurdles.

Integrating Silos

One of the primary obstacles is the lack of a universal standard. Currently, buy-side and sell-side systems are being developed in silos. "I feel like both sides need to talk while they’re building their platforms," Karmali stated. "At the moment, everything gets built in silos, and then we have to figure out how to connect it in the middle."

The "Common Language" Problem

For an agent to function, it needs to understand the nuance of inventory quality, pricing, and authorization. When a buyer’s agent approaches a seller’s agent, there must be a common vocabulary to negotiate a deal. Currently, there are multiple competing industry frameworks, and no consensus exists on:

1. Guardrails: How to prevent agents from gaming the system or circumventing established trading practices.
2. Authorization: How a buyer’s agent verifies that the seller’s agent has the authority to represent specific, high-quality inventory.
3. Liability: Determining accountability when an AI agent misinterprets a brief and commits significant budget to the wrong inventory.

Why CNN is Betting on AI

CNN International’s motivation is clear: they want to reclaim control over how their inventory is described and sold. In the current programmatic ecosystem, publishers are often at the mercy of opaque algorithms that undervalue premium content.

By utilizing a "clean metadata layer" built on human-verified, fact-checked content, CNN aims to package its inventory in a way that AI agents can easily understand and prioritize. "We know we have a clean environment with human-made fact-check content that has a clean metadata layer," Karmali explained. This is the "prize"—the ability to ensure that premium journalism is not lost in the "noise" of the programmatic long-tail.

However, the efficacy of this strategy depends entirely on the buyer’s agent. If the buyer’s agent cannot accurately interpret the metadata or the value of the "provenance" CNN offers, the technical effort may yield diminishing returns. As it stands, the industry is still debating what an "accurate description" of premium ad inventory actually looks like in the age of AI.

Market Perspectives: Efficiency vs. Delivery

The industry’s reaction to agentic trading is mixed. While some view it as a necessary evolution, others warn against simply automating the inefficiencies of the past.

Ravi Patel, CEO of the ad tech startup SWYM, acknowledges the benefits of efficiency but offers a cautionary note. "Seller agents are a genuine efficiency gain for direct sales. They take cost out of the RFP-to-trafficking grind," Patel said. "But they automate the deal, not the delivery. Performance is decided impression by impression, in milliseconds, on bid-request signals. Connect those two layers and you’ve got something. Otherwise, you’ve automated the paperwork, not the performance."

This sentiment underscores the distinction between "trading" (the contract) and "execution" (the campaign delivery). If CNN’s agents can bridge this gap, they could revolutionize the market. If they only automate the paperwork, they may find that advertisers—who are currently wary of adding more complexity to their stacks—remain unconvinced.

Implications for Future Revenue

When asked about the financial impact, Karmali remained circumspect. He declined to comment on whether agents would definitively shift the balance of power toward publishers, noting that it is "too early to say."

In terms of revenue contribution, the projections are wide-ranging. Agent-led deals could start as a modest 5% of CNN International’s ad business, with the potential to grow to 50% in the long term. "There’s still a path to go," he added, acknowledging that the transition is a marathon, not a sprint.

The Broader Context: CNN’s Multi-Platform Evolution

While the agentic experiment captures the headlines, it is only one piece of a broader, aggressive growth strategy at CNN. The network is currently diversifying its revenue streams through several key initiatives:

• FAST Channels: CNN’s "Headlines" channel, which utilizes clips from its online operations, is now active in 50 countries, proving the viability of scalable, automated content distribution.
• Subscription Growth: The CNN All Access service, which launched in the U.K. and Canada on October 20, 2026, integrates advertising from day one, signaling a shift toward a hybrid subscription-advertising model.
• Global Perspectives: This editorially led events series is the "antithesis of agentic." By hosting high-level, sponsor-funded gatherings from Davos to Bangkok, CNN is leaning into the premium, human-centric, and high-touch nature of its brand.

Conclusion

CNN International is effectively running a two-track strategy: one track is high-touch, human-led, and event-based, while the other is highly automated, agent-driven, and data-centric. By pushing into the world of agentic workflows, the network is attempting to future-proof its business against a digital landscape that is rapidly becoming automated.

The success of this endeavor will hinge on the industry’s ability to move beyond silos and establish a shared, common language for AI-to-AI trading. Until that happens, CNN’s technology will remain a sophisticated internal tool waiting for the market to catch up. Whether this leads to a new era of publisher-led pricing power or simply a faster way to process the status quo remains the industry’s most significant unanswered question heading into 2027.