Grigoriy MisilyukThe foundation of programmatic advertising is the OpenRTB protocol. Will this change? No — OpenRTB remains essential. But are new approaches emerging alongside it? Yes, and that is what this article explores.
Artificial intelligence is becoming a core operational layer across digital markets. Automation and machine-driven decision-making are no longer experimental — they are increasingly embedded in everyday workflows. Programmatic advertising is no exception.
At the same time, the traditional open-auction model is showing structural limitations. While OpenRTB remains the foundation of programmatic trading, open marketplaces are highly fragmented and increasingly inefficient for premium inventory. For many publishers, they no longer represent the main source of sustainable revenue growth.
As a result, large publishers and sales organizations have shifted toward direct and curated deals. These models provide better pricing control, stronger demand quality, and more predictable outcomes. However, they also rely on lengthy negotiations, custom workflows, and ongoing operational coordination — making them difficult to scale.
The industry is now experimenting with agentic trading models, where AI agents negotiate, activate, and manage deals based on predefined business rules rather than human-driven processes. Standards such as AdCP, ARTF, and UCP are emerging to support this shift. They are not designed to replace OpenRTB, but to complement it by enabling more structured, automated, and scalable forms of trading.
What these new frameworks actually change
If you’re encountering these standards for the first time, it’s helpful to think of them as new ways to organize direct and PMP deals using agentic frameworks.
Long story short: despite being positioned as “alternatives” or “next steps,” AdCP, ARTF, and UCP are not designed to replace OpenRTB. They are new technical standards designed to support emerging trading models, including agent-to-agent execution, intent-based deal management, and workflow-driven direct trading.
Programmatic trading models spectrum
Together, these frameworks address the growing demand for trading models that sit between open auctions and fully manual, direct deals — more controlled than the open market but more flexible than traditional insertion orders. They shift decision-making away from individual bid requests and toward shared rules, workflows, and platform-level logic. As a result, fewer decisions are made on an impression-by-impression basis, and more are defined in advance.
Ad Context Protocol (AdCP) is a standard for structuring how agents and platforms express advertising intent. It enables negotiation and campaign execution through predefined workflows.
Agentic RTB Framework (ARTF) is a standard for moving execution logic closer to the core RTB infrastructure. It enables services and decision logic to run within the platform, reducing reliance on external calls and on fragile, millisecond-level orchestration.
User Context Protocol (UCP) is a standard for exchanging user and contextual signals between systems in a compact, structured format. These representations replace large, verbose data payloads.
Taken together, these approaches do not “simplify the auction.” Instead, they create an alternative to the sole decision point in programmatic trading. OpenRTB remains foundational, but it is increasingly complemented by mechanisms that enable participants to define how trading operates.
Why the sell-side is experimenting now
The main reason is that it can be profitable, helping you both earn more while spending less on infrastructure.
The roots are in the complexity of programmatic: OpenRTB is built for open, fragmented markets. Like any solid structure, it has inertia. It lies in the large number of mandatory fields that must be transmitted during trading. As a result, every additional query increases infrastructure costs, limiting traffic scaling.
Agentic-based and agentic-assisted trading address the interests of the sell-side from multiple perspectives:
Infrastructure cost and scalability. Agentic approaches allow platforms to reduce per-impression computational load by shifting decision-making to higher-level rules and workflows. This enables traffic and revenue growth without linear increases in infrastructure spending.
Rising operational complexity. As the programmatic stack expands, managing integrations, deal types, and optimization logic becomes increasingly difficult. Agentic systems centralize execution logic, allowing teams to scale operations without proportional increases in configuration, maintenance, and support effort.
Efficiency gains through agentic-based systems. Agentic trading frameworks are better aligned with machine learning systems that operate on patterns, intent, and aggregated signals. This enables more consistent optimization at scale, without relying on constant manual tuning and impression-level intervention.
Greater control and predictability in trading. By moving part of the decision logic outside the millisecond auction, sell-side platforms gain stronger control over how inventory is sold and to whom. This improves revenue predictability and reduces dependence on highly dynamic auction behavior.
Time savings by reducing direct negotiations. Standardized workflows and agent-driven interactions reduce the need for ongoing one-on-one negotiations. This allows publishers and platforms to scale partnerships and deal execution without proportional increases in operational overhead.
These are the main drivers of current experimentation on the sell side. Yet there are additional considerations, as each new technology path significantly expands monetization opportunities.
How these changes reshape sell-side strategy
First, without owning or controlling infrastructure, agentic trading is not realistically achievable. Self-serve and fully managed platforms rarely provide the level of control required for this type of experimentation. In some cases, publishers may access agentic capabilities through existing SSP partners. But long-term flexibility still depends on having direct influence over how trading logic is implemented.
For publishers and sell-side platforms, this signals a shift from relying on external solutions toward owning the systems that define how trading actually works.
Not every player needs to move at the same speed — but delaying experimentation carries real opportunity cost. For some, these changes will unlock near-term advantages by enabling faster testing of new deal types and execution models. For others, the impact may remain gradual. The key differentiator is not scale, but whether a company can change how it trades without rebuilding its stack.
This is where infrastructure strategy becomes decisive. Agent-to-agent workflows can be launched in different ways — through existing SSP partners, shared platforms, or dedicated environments built in collaboration with technology providers. For many publishers, this will be the most practical entry point.
At the same time, having deeper control over the trading environment creates long-term advantages. Platforms that consolidate execution, reporting, and optimization in one place make it easier to test new models, compare results, and scale what works. When combined with an embedded AdCP sales agent, this setup enables teams to iterate faster—without constant renegotiation or fragmented tooling.
The strategic value here is not ownership for its own sake. It’s about having sufficient architectural flexibility to experiment, learn, and adapt as agentic trading matures.
What matters most for monetization strategy now
The real challenge is not predicting which standard will prevail. The more important task is recognizing trends indicating structural change in who controls the trading logic and where decisions are made.
The signals are already there: closed, managed trading models are becoming more common, and decision-making is moving from individual auctions to platform-level rules. At the same time, platforms increasingly need to integrate with agentic and ML-driven systems.
In discussions with technology vendors, feature checklists are increasingly irrelevant. The focus should be on operational adaptability: How easily can trading rules be changed? Where is execution logic actually enforced? Does the platform maintain control over data and signals as new models are introduced? These answers determine whether experimentation is realistic or purely theoretical.
In this environment, flexibility matters more than choosing the “right” standard. Platforms that allow testing and continuous, low-risk experimentation create a strategic advantage regardless of how the ecosystem evolves. By contrast, rigid systems risk becoming constraints precisely when the market begins to move faster.
The takeaway is simple: choices sell-side leaders make today either increase their ability to adapt tomorrow or lock them into operating models that are hard to change. In the near term, architectural flexibility — not protocol selection — is the defining strategic factor.
