For years, the advertising industry has evolved around a constant promise: more automation, more data, more efficiency. From programmatic buying to advanced algorithmic optimization models, the goal has been to reduce operational friction and increase precision.
But the change that is now beginning to take shape is not incremental. It is structural.
The industry is facing what will likely be its biggest transformation since the arrival of programmatic advertising: the emergence of agentic advertising. A model that no longer limits itself to automating tasks, but instead shifts part of the decision-making process to systems capable of reasoning, executing, and optimizing autonomously.
The difference is profound. Until now, technology merely responded. Now, it is beginning to act.
Until now, Artificial Intelligence has mainly been used to create copy, images, or ad variations for A/B testing, analyze data, summarize reports, recommend audiences, and automate rules.
It is a major step forward, but it still remains assistive in nature: the human asks, and the tool responds.
In the agentic model, the system operates more like an “operational team” of agents. You simply provide an objective, and the agent can break that objective down into tasks, reason based on signals, execute actions such as launching advertising campaigns or adjusting creatives and budgets, and continuously adapt in pursuit of the desired outcomes in order to achieve the proposed goal.
Instead of having a person manually working across a DSP, an analytics tool, a creative generator, a CDP (Customer Data Platform), and a reporting spreadsheet, an agentic system can coordinate many of these functions. It can act as an intelligent layer connecting data, platforms, inventory, creatives, and business objectives.
To understand this shift, we need to place it within the third wave of Artificial Intelligence:
In this new paradigm, AI stops being merely a support tool and becomes an orchestration layer. While traditional programmatic advertising optimizes isolated variables, the agentic model manages missions. A brief is no longer a closed technical instruction; it becomes an objective entrusted to a system that preserves the campaign’s “intent” over time, continuously adjusting strategy in real time.
To evolve in this direction, the IAB Tech Lab is working on AAMP (Agentic Advertising Management Protocols), an initiative aimed at standardizing how agents can operate within digital advertising. It is not starting from scratch: it builds on existing standards such as OpenRTB, AdCOM, OpenDirect, VAST, Deals API, OMID/OM SDK, GPP, and TCF. In other words, the idea is not to replace the entire advertising infrastructure, but to “agentify” it: enabling buying, selling, measurement, and management agents to interact with systems that already exist.
In one sentence: Agentic Advertising is advertising managed by intelligent agents that not only generate ideas or creatives, but also make operational decisions to achieve a business objective.
Historically, the adtech ecosystem has been built on a fragmented logic: disconnected platforms, distributed data, and teams specialized by function (planning, buying, measurement).
For the advertising industry, Agentic Advertising could have an impact similar to the one programmatic advertising once had, but at the operational layer, rather than at the technical layer initially defined through OpenRTB.
Agentic advertising introduces a cross-functional intelligence layer capable of coordinating decisions across heterogeneous systems. The capabilities of these agents make it possible to:
IAB Europe has framed it precisely in these terms: if AI becomes integrated across the entire advertising value chain, the industry will need shared frameworks that guarantee interoperability, transparency, accountability, and trust between buying, selling, and measurement.
In this scenario, the critical infrastructure is no longer just inventory, but the quality of APIs and data interoperability, since an agent’s ability to operate depends on them.
The deepest transformation is not technological, but organizational.
Although there are no official figures, several studies and industry analyses estimate that the rise of programmatic advertising (2012–2020) already reduced between 30% and 60% of manual trafficking and operational management tasks. The agentic era, however, now impacts the intermediate layers of supervision.
If yesterday’s professional manually adjusted bids and corrected budget deviations, tomorrow’s professional will move toward higher-value strategic functions:
The value no longer lies in being the fastest at execution, but in being the most precise at defining criteria. The campaign manager becomes a manager of autonomous systems.
This shift will also have labor implications. There will be less value in mechanical execution and greater value in profiles capable of combining strategy, data, creativity, editorial judgment, ethics, privacy, and technical expertise. The differentiating professional will no longer be the one who knows how to manually launch a campaign, but the one who knows how to manage a system that launches, measures, and learns.
But there are also risks. If too much responsibility is delegated too early, teams may lose control. If agents optimize only for short-term outcomes, they may damage brand value. If low-quality signals are used, errors may increase. And if nobody understands why the system made a decision, efficiency turns into a black box.
That is why the concept of human-in-the-loop will remain essential. Humans may not be involved in every click, but they will still need to be involved in system design, controls, sensitive approvals, and the strategic interpretation of results.
While this model applies across the entire digital ecosystem, it finds its strongest justification in Connected TV (CTV). Because of its scarcity of premium inventory, fragmentation, and high operational complexity, CTV is the perfect testing ground for agentic AI.
An agent applied to CTV could help determine which publishers to invest in, which deals to prioritize, which audiences to activate, what maximum frequency to apply, which creatives perform best by device or environment, and how to distribute budgets across FAST, AVOD, broadcaster apps, OEM platforms, or premium programmatic inventory.
It is important to emphasize that an agent cannot behave as though it were optimizing open display inventory. In CTV there are ad pods, breaks, spot duration constraints, competitive separation rules, frequency controls, latency considerations, SSAI, OMID/OM SDK measurement where applicable, VAST standards, creative compatibility requirements, and publisher quality policies. Automation must respect both the content and the viewing experience.
Agentic advertising represents the shift from an industry centered on transactions to one centered on orchestration. Television, historically based on manual planning and estimations, is evolving into an addressable algorithmic system where every impression can be optimized.
If programmatic advertising was the layer that digitized media buying, AI agents are the infrastructure that will make it truly autonomous. This future is consolidating on the largest screen in the home, where today’s operational complexity makes the intervention of intelligent agents not just a competitive advantage, but an absolute necessity for scaling the channel.
At tvads we has a professional team able to advise you on this field and and guide you in any area of your streaming advertising business, advising you or even operating it on your behalf if necessary
All author posts