WPP Media Builds AI Buyer Agent for Video Ads

WPP Media, the media arm of advertising holding giant WPP, is developing an AI buyer agent for video as part of a broader Agentic Standards Initiative aimed at bringing autonomous, machine-to-machine negotiation into programmatic media buying. The effort signals one of the clearest moves yet by a major buy-side player to operationalize "agentic" workflows inside the video and CTV ad-buying stack.

For an ad-tech audience, the headline is less about a single product and more about what it represents: the early scaffolding of an agent-driven layer sitting on top of — or eventually inside — the programmatic plumbing that already powers DSP-to-SSP transactions over OpenRTB.

What an AI buyer agent actually does

In practical terms, a buyer agent is software that can interpret a campaign brief, translate strategic objectives into buying parameters, and then execute and optimize against inventory with limited human intervention. Where a traditional trader manually configures line items, bid strategies, frequency caps, and supply paths inside a DSP like DV360 or The Trade Desk, an agent is intended to reason over those decisions continuously and act on them.

For video specifically, that decision space is dense. A buyer agent operating in CTV and online video must account for ad pod construction, VAST response handling, competitive separation, frequency management across fragmented apps and devices, and supply-path quality — all while reconciling reach and outcome goals. Automating that reliably requires structured, machine-readable signals on both sides of the transaction, which is exactly where standards become the gating factor.

Why "standards" is the operative word

The Agentic Standards Initiative framing matters because agents are only as useful as the interfaces they can talk to. An AI buyer agent that can only operate inside one DSP or one walled garden delivers limited value. The strategic ambition is interoperability: agents on the buy side negotiating with agents (or standardized endpoints) on the sell side, exchanging intent, constraints, and inventory availability in a consistent format.

This is conceptually adjacent to the work the industry already relies on — OpenRTB for bid requests and responses, ads.txt/app-ads.txt and sellers.json for supply-chain transparency, and VAST for video ad delivery. An agentic layer would not replace those; it would sit above them, using standardized protocols to let autonomous systems coordinate. Without shared schemas for how agents express objectives and how sellers expose inventory and pricing, every integration becomes bespoke and the ecosystem fragments.

Implications for the sell side and CTV

If buy-side agents become a meaningful share of demand, SSPs, exchanges, and CTV platforms will face pressure to expose machine-negotiable interfaces. That could reshape how video supply is packaged and described — pushing toward richer, more structured metadata on content, audience, pod position, and outcome measurement so that agents can evaluate inventory programmatically rather than relying on human-curated deal IDs.

CTV is a natural proving ground. The channel already concentrates spend, carries high CPMs, and suffers from fragmentation across publishers, apps, and devices — precisely the kind of complexity where automated optimization can theoretically outperform manual trading. But CTV also has thinner standardization in areas like identity, frequency capping across publishers, and cross-platform measurement, which constrains how much an agent can reliably reason over today.

The open questions

Several practical issues will determine whether this becomes infrastructure or marketing. First, governance: who defines the agentic standards, and will sell-side platforms and IAB Tech Lab–style bodies converge on a shared spec, or will competing protocols emerge? Second, transparency and auditability: autonomous buying raises real questions about how decisions are logged, how brand-safety and supply-quality controls are enforced, and how advertisers verify that an agent optimized toward the stated objective rather than a proxy metric. Third, fraud and IVT exposure — automated, high-velocity buying without robust verification could amplify exposure to invalid traffic if measurement signals aren't embedded into the agent's decision loop.

For ad ops teams and publishers, the near-term takeaway is to watch how these standards develop and how they intersect with existing programmatic infrastructure. An agentic buying layer that respects OpenRTB, supply-chain transparency, and standardized video delivery would integrate far more smoothly than one that tries to route around them.

WPP Media's involvement gives the effort weight — buy-side scale is what forces sell-side adoption. Whether the Agentic Standards Initiative produces genuinely interoperable protocols or remains a buyer-controlled stack will be the signal worth tracking.

Stay on top of video ad serving and programmatic. Follow Adelerate.