AI agents have bought roughly $1.2 million in media over the last three months for "very large" brands through the platform Apostra, according to the company — a small figure that signals advertisers are now committing real budgets to agentic buying.

September has grounded the agentic ad-buying pitch in reality: small budgets, tangible results, and humans still watching closely. The promise remains that agents plan, buy, and optimize media, relieving humans of the grunt work. The current deployments show the technology is moving beyond experimentation and into execution across the advertising workflow — but it is still a long way from running advertising on its own.

What can agents actually do today?

Apostra, which has pivoted from sustainability to agentic media buying, lets AI agents discover, negotiate, and approve deals, then execute them across digital publishers, linear TV, and in-store audio. Ad buyers can use the platform to connect with other agents through agentic connectors for major platforms, managing campaigns across Meta, Google, Pinterest, Reddit, Spotify, and ChatGPT.

Other adtech players are assembling similar systems:

• The Trade Desk is building a central AI conversational interface that will route media buyers to a growing roster of specialized agents.
• PubMatic AgenticOS launched in January.
• Magnite has introduced a buyer agent and an agentic orchestration layer spanning buy- and sell-side functions.
• Kargo introduced Project Kera, an AI-backed planning and buying interface, in Spring.

The agentic push in September extended beyond media buying. Runway, an AI video startup, entered the ad business with Runway Ads, an AI agent that creates variations of existing ads, localizes them for Meta, Google, and TikTok, and then pulls performance data. At Dreamforce, Southwest Airlines said its AI agents will soon handle more complex tasks, including completing a flight ticket purchase. WPP opened Devon's Point, a hub that uses AI-powered tools to automate parts of the development and production process.

What results are advertisers seeing?

Red-light therapy brand Rouge Care credits AI agents with a fivefold return on ad spend. The brand spent $25,000 on CTV ads and drove over $125,000 in attributable sales from an awareness campaign facilitated through PubMatic's AgenticOS in the second quarter of 2026.

Still, no one is doing this at scale. Highly autonomous deployments represent less than 5% of all marketing use cases in 2025, according to Gartner. The firm predicts that by 2028, only 15% of agentic AI deployments will be highly autonomous.

For now, the question is not whether advertisers will put money behind agentic buying, but how far that spending can scale — and what infrastructure will ultimately support it.

Who sets the rules agents speak?

Apostra uses AdCP, a protocol developed by a group of adtech companies including Scope3 that provides a common language for agents to interact. It is one of a handful of protocols in the ad industry, alongside IAB Tech Lab's AAMP (Agentic Advertising Management Protocols) and Amazon's Model Context Protocol.

The rise of multiple protocols could create fragmentation issues and make agentic advertising harder to scale. If buyers and publishers back different standards, the systems will need a way to communicate — or risk adding another interoperability layer to an industry that is already notoriously fragmented.

Apostra pitches itself as a neutral meeting point for agents working with brands and publishers. It does not take a cut of media spend, as ad networks might. It positions itself as software-as-a-service, with a fee that does not increase with the amount of ad dollars flowing through it. Andrew Frank, VP, distinguished analyst at Gartner, said that model could give advertisers a reason to favor a neutral coordination layer over platforms like Amazon or Meta.

Who is accountable when agents go wrong?

Even if the plumbing works, hard questions remain about who is responsible for the decisions agents make.

"The problem is agents are supposed to make strategic decisions about budget allocations," Frank said. "If agents reallocate, say $10 million from TV to creator inventory, and performance drops, who's accountable? Agentic buying introduces this whole new decision layer whose recommendations may not be fully transparent."

AI might reveal some of its reasoning behind a decision, but that reasoning may still fail. For brands, that risk could keep budgets in the experimental bucket.

Handing over control will be particularly difficult in premium advertising, Frank said, where large deals remain heavily relationship-driven and publishers and buyers have spent years building their own infrastructure. Until accountability and interoperability questions get answers, agentic buying will grow — but from a very small base.