The promise of agentic AI in advertising is moving from demos to live media buying. Magnite and Amnet France have now put that proposition to a real-world test, using AI agents to plan and activate a premium connected TV campaign for an automotive advertiser. The campaign reportedly cut setup time by about 70% while producing a 95% video view-through rate, offering an early example of how agent-to-agent workflows could change the mechanics of programmatic CTV buying.

For years, programmatic advertising has automated much of the transaction between advertisers and media owners. But the work surrounding those transactions has remained surprisingly manual.

Media teams still spend significant time translating campaign briefs into platform settings, finding appropriate inventory, identifying publishers, structuring deals and monitoring opportunities across fragmented supply. Magnite’s latest experiment suggests that the next phase of automation could involve AI agents handling more of that operational layer.

The company said its first agentic campaign in EMEA was conducted with Amnet France, the trading desk operated by Dentsu, for a leading automotive manufacturer. Amnet used natural-language prompts with Magnite’s buyer agent to build and activate a video campaign through ClearLine.

The important part of the workflow was what happened between the buyer and seller sides.

Through Magnite Orchestration, the buyer agent communicated with Magnite’s seller agent to identify and activate premium CTV inventory that matched the campaign objectives. Instead of manually working through campaign configuration, publisher discovery and deal construction, the agency team could concentrate more heavily on campaign strategy.

According to Magnite, the process reduced campaign setup time by approximately 70%. The campaign also generated a 95% video view-through rate, while the agent identified inventory and optimization opportunities that might not have surfaced through conventional workflows.

That distinction matters. This is not simply another generative AI interface placed on top of an advertising platform.

The emerging agentic model attempts to connect an instruction with the systems capable of acting on it. A buyer could express a campaign objective in natural language, while software agents interpret that objective, search available supply, evaluate potential opportunities and execute transactions within defined parameters.

Magnite has been building toward that model throughout 2026. In June, the company introduced Magnite Orchestration as a coordination layer connecting buyer agents with its seller agent. The platform is designed to allow buyers to discover and activate premium omnichannel inventory while also giving publishers and data providers ways to make audiences and inventory available to AI-driven buying systems.

The EMEA campaign provides a more tangible demonstration of that architecture.

For agencies, the potential benefit is less about eliminating media buyers and more about changing where their time is spent. If repetitive execution tasks can be delegated to software agents, trading teams could devote more attention to audience strategy, supply quality, creative decisions, measurement and optimization.

That could become increasingly relevant as CTV buying becomes more sophisticated. Premium streaming inventory is distributed across broadcasters, streaming services, connected TV platforms, apps and other digital environments. Automated decision-making can potentially help buyers navigate that complexity, but it also introduces a new question: how much control should an advertiser give an AI system over media decisions?

The industry is already moving toward that question. Gartner reported in August 2026 that more than 70% of global advertising spend is expected to flow through self-service advertising platforms where AI materially influences media buying, cost or outcomes by 2028.

Gartner has also projected that by the fourth quarter of 2026, 60% of marketing organizations with more than $1 billion in revenue will have piloted or deployed complex workflows incorporating AI agents and multiple data sources.

Those forecasts help explain why Magnite’s experiment matters beyond one automotive campaign. The advertising industry is moving from AI-assisted execution toward AI-mediated execution, where software increasingly participates in the decisions that determine which inventory is purchased and how campaigns are optimized.

Still, the 70% efficiency improvement should be treated as an individual campaign result rather than proof that agentic buying will deliver the same gains universally. Campaign complexity, inventory availability, data quality, agency controls and platform integrations will all influence the outcome.

There is also a transparency challenge. When an AI agent selects inventory or recommends optimization opportunities, advertisers need to understand why those decisions were made and whether the system is prioritizing business objectives, media quality and independent measurement appropriately. Gartner has warned that deeper AI involvement in advertising can make media decisions harder for marketers to explain and justify.

That makes interoperability and governance just as important as automation.

Magnite’s approach is notable because it connects the buyer and seller sides rather than treating AI as an isolated assistant. Its broader Orchestration strategy is intended to let buyers connect their own agents to Magnite’s seller-side infrastructure, creating a framework where different systems can participate in the same advertising workflow.

For enterprise advertisers and agencies, the practical question will therefore be less about whether they can deploy an AI agent and more about where that agent sits within the existing media stack.

If agentic buying develops as Magnite and other advertising technology companies expect, campaign briefs could increasingly become machine-readable instructions, while DSPs, SSPs, data platforms and measurement systems become the execution layer behind those instructions.

The Amnet France campaign is an early signal of that shift. Its most important result may not be the 70% reduction in setup time, but the demonstration that an AI agent can move from understanding a media objective to interacting with advertising supply and executing a CTV campaign.

That is a considerably bigger change than simply adding a chatbot to a DSP.

Market Landscape

Agentic advertising is emerging as programmatic platforms look to automate more than bidding and optimization. Magnite’s June launch of Orchestration positioned buyer and seller agents as participants in the same transaction environment, while its April AI expansion included buyer-agent capabilities alongside AI tools for media owners.

The competitive landscape includes established ecosystems such as Google and Amazon, enterprise marketing platforms from companies such as Salesforce and Adobe, and the broader advertising infrastructure provided by DSPs, SSPs and CTV marketplaces. NVIDIA’s AI infrastructure is also part of the underlying technology shift, although the advertising workflow itself remains controlled by specialized media platforms.

For agencies, the likely battleground will be interoperability, inventory quality, identity, measurement and control—not simply which vendor has the most capable AI model.

Top Insights

• Magnite and Amnet France used buyer and seller AI agents to automate premium CTV campaign activation, reducing manual setup work for agency trading teams.
• The EMEA campaign reportedly cut setup time by about 70% and delivered a 95% video view-through rate for an automotive advertiser.
• Magnite Orchestration connects buyer agents with seller-side inventory intelligence, signaling a shift from AI assistance toward agent-driven programmatic execution.
• Enterprise advertisers could gain operational efficiency, but governance, transparency, measurement and human oversight will remain critical as AI influences media decisions.
• Gartner expects AI-influenced self-service advertising platforms to account for more than 70% of global ad spend by 2028, accelerating pressure on media teams to adapt.

Get in touch with our Adtech experts.