By Thomas Bernal, SVP, Offer Strategy & Execution, Ogury
The ad tech ecosystem is notoriously fragmented, with thousands of players each operating their own technologies, datasets, and business models. This is one of the biggest barriers preventing the open web from performing as seamlessly as walled gardens. In such a fragmented landscape, a growing number of open web players are calling for greater interoperability. The question isn’t about whether or not it’s realistic (because it is); it’s about what form interoperability could take and how quickly industry players can align around it.
Recent collaborations are showing that there is more to be gained through cooperation than competition: Amazon has partnered with Roku and Netflix, while Google has teamed up with Criteo, showing that even the biggest fish in the media pond are willing to dismantle silos for mutual gain.
Let’s look at what’s helping and hindering interoperability today, why achieving it is vital for the open web’s future, and whether AI can help us get there.
Will consolidation open doors or put up walls?
Open web advertising, and indeed the wider media ecosystem, is in a period of consolidation. Much of that trend is driven by financial headwinds that have made many companies ripe for acquisition and made cost control more important than ever.
Consolidation can itself solve interoperability by reducing hops between supply and demand and limiting third-party involvement. This can be seen in SSPs introducing buying platforms and DSPs moving towards direct access to supply, both aiming to create a leaner, more efficient supply chain.
Does this mean the open web will splinter into mini walled gardens, each ruling its own closed loop? Most likely, no. Walled gardens skirt the need for interoperability because their supply is in such high demand that they can own and operate trading pipes end-to-end. The open web, by contrast, will never have a single player with that dominance. Publishers will always seek multiple demand paths, and advertisers multiple supply paths.
Those paths may shrink in number, but they must remain broadly compatible, as no single party can dictate the rules of the market. For independent ad tech players who haven’t been swept up in the consolidation wave, interoperability isn’t optional, it’s a matter of survival.
Why interoperability is harder than it looks
Achieving interoperability isn’t as straightforward as it may seem, and attempts to solve fragmentation can sometimes lead to new inefficiencies. Alternative IDs are a prime example. They were designed to unify the ecosystem, but in practice, every translation adds friction, another tech layer, another vendor, and another fee. These types of cost can make access to first-party data prohibitively expensive and slow down, or even derail, collective initiatives.
Supply-side curation provides another avenue for activating first-party data; in this case, from publishers. Publishers are naturally cautious about giving up too much control to intermediaries in a supply chain that has often commodified their inventory. By packaging their first-party data themselves, publishers can preserve control while adding value through curated deals to advertisers. However, curation still raises questions around control and how much value is truly added (and who extracts that value) once all fees are accounted for.
AI demands interoperability and can deliver it too
AI is intensifying the need for interoperability. On one hand, it opens new paths to overcome fragmentation, for example, it’s now relatively easy to deploy AI agents that act as a “meta DSP” that can bridge multiple buying platforms. At the same time, AI also reinforces existing barriers, such as siloed or poorly taxonomised data, which make it harder to automate complex operations at scale or leverage the predictive power of machine learning.
With many interoperability challenges being data-scale issues, AI also serves as a solution to these roadblocks, translating disparate data sets into common taxonomies and expanding seeds of audience data into market-ready segments that can reach across the entire open web, and beyond.
Loftier expectations around agentic advertising must be tempered. In theory, agentic could manage the entire advertising chain, translating between platforms and solving interoperability in one fell swoop. In reality, AI agents are better suited to repetitive, resource-heavy tasks like mapping inventory to audiences, optimising supply paths, or adjusting campaign bidding strategy..
Programmatic trading runs in milliseconds, millions of times a second. Agent-to-agent communication adds too much server load and latency, making real-time decisioning unworkable at scale. Unless quantum computing arrives sooner than expected, a fully agent-run programmatic chain will remain science fiction.
Beyond feasibility lies the money question: even if technically possible, who will bear the added computational costs of an agentic ecosystem, and what value would it add to justify those costs?
Agentic AI: from hype to implementation
Today, most agentic AI in programmatic operates within siloed, closed-loop systems or through existing APIs rather than through native AI-to-AI communication. Internal agents are built by companies to manage processes within their own platforms; for instance, a DSP optimising bids or an SSP improving yield. These agents are not speaking directly to third-party systems.
When cross-platform interaction is needed, it typically happens via APIs, which were designed for structured data requests and responses, not for the autonomous, goal-driven dialogue between agents.
The industry has a precedent for creating shared standards like OpenRTB or OpenDirect, but these take years of multi-stakeholder collaboration. Early efforts like the Model Context Protocol (MCP) and A2A initiatives are promising steps, but a widely adopted protocol for agentic AI that ensures trust, governance, and transparency remains a long-term project. Until then, agentic AI will remain constrained to internal stacks or legacy API connections rather than the seamless interoperability many envision.
In the here and now, the challenges and opportunities around interoperability will continue to shape open web advertising’s supply chain and the priorities of those who operate within it. Those who can turn this complexity into a competitive edge will lead the open web’s next evolution, not by controlling the most data, but by connecting most effectively. Interoperability must become the default if the ecosystem is to unlock its full potential and finally stand on equal footing with walled gardens.
