The world of programmatic advertising has long been plagued by complexity. Buyers navigate a labyrinth of dashboards, vendor integrations, and manual campaign adjustments, a process that’s often inefficient, and opaque. Now, PubMatic is betting that artificial intelligence, specifically a new generation of AI “agents,” can finally streamline this convoluted system and unlock greater value for advertisers. The company’s push, built on its AgenticOS infrastructure and a new partnership with Untapped Growth, a collective of 20 independent agencies, aims to automate much of the heavy lifting in digital ad buying.
At the heart of this effort is a shift away from traditional, human-orchestrated workflows. PubMatic’s AgenticOS, developed in collaboration with NVIDIA, replaces the need to toggle between multiple platforms with a network of specialized AI agents. These agents handle tasks ranging from audience discovery to campaign execution and optimization, all within a unified system. The goal is to create a more fluid and responsive process, reducing wasted time and improving campaign performance. This isn’t simply about automation; it’s about fundamentally changing how programmatic advertising operates.
Simplifying a Fragmented Ecosystem
The current programmatic landscape is notoriously fragmented. Buyers typically interact with demand-side platforms (DSPs) that connect to supply-side platforms (SSPs), which in turn connect to publishers. This multi-layered structure adds complexity and cost. According to Harry Tong, director of sales engineering at PubMatic, AgenticOS is designed to collapse this traditional supply chain. “The legacy view…where there is a buyer who logs into a DSP who connects to an SSP who connects to publishers on the back end—that supply chain is collapsing,” Tong explained. “SSP, DSP terms are already a bit obsolete.”
This simplification is particularly appealing to independent agencies, which often lack the resources and scale of larger holding companies. Untapped Growth, representing these agencies, is leveraging the PubMatic partnership to offer its members access to sophisticated AI-powered tools. Brendan Clifford, co-founder and managing partner at Untapped Growth, described how their agencies can now input a media plan into a large language model (LLM), such as Claude, and execute transactions directly with SSPs, bypassing the traditional DSP layer. This, he argues, cuts down on the “tech tax and the operating tax” associated with using a DSP.
“Traditional buying approach for indies, where they are taking a brief from a client, and put together a plan, contracting with an SSP and a DSP, and all of that has been condensed into one remarkably efficient LLM prompt, reducing the time to market of that solution and ensuring the best possible cost efficiency to the highest premium inventory,” Clifford said.
Early Results and the Promise of Agentic Systems
While the concept of agentic systems in programmatic advertising is still relatively new, early results are promising. PubMatic reports that a third agency testing AgenticOS saw an 82% reduction in buy-side costs and a nearly 90% reduction in campaign setup time. These figures suggest a significant improvement in efficiency and resource allocation. Two additional agencies within Untapped Growth are slated to begin using the system shortly, providing further real-world data points.
The integration of AI agents isn’t just about speed and cost savings; it’s also about unlocking new levels of personalization and optimization. By automating routine tasks, advertisers can free up their teams to focus on strategic initiatives and creative development. The ability to quickly analyze data and adjust campaigns in real-time allows for more targeted and effective advertising.
The Future of Programmatic: Beyond DSPs and SSPs?
The long-term implications of this shift are significant. If PubMatic’s vision takes hold, the traditional distinctions between DSPs, SSPs, and other players in the programmatic ecosystem could begin to blur. The rise of agentic systems may lead to a more streamlined and integrated marketplace, where AI agents act as intermediaries, negotiating deals and optimizing campaigns on behalf of advertisers. This could ultimately empower buyers with greater control and transparency.
Though, questions remain. Whether agentic systems can *fully* replace human orchestration, or simply shift where that orchestration happens, is still an open debate. The success of this approach will depend on the ability of AI agents to adapt to changing market conditions and deliver consistent results. The industry is watching closely to see if PubMatic’s bet on AI will pay off.
PubMatic’s AgenticOS is currently available to select partners, with broader rollout planned throughout the year. The company is continuing to invest in AI research and development, with a focus on expanding the capabilities of its agentic infrastructure. The next key milestone will be the full integration of AgenticOS with additional DSPs and SSPs, further expanding its reach and impact on the programmatic landscape.
What do you think about the potential of AI to transform programmatic advertising? Share your thoughts in the comments below.
