Optable and PubMatic partner to connect publisher first-party data with agentic media buying workflows via AgenticOS, advancing the Advertising Context Protocol in programmatic.
Optable and PubMatic this week announced a partnership that embeds Optable's Audience Agent directly into PubMatic's AgenticOS, enabling buyers to autonomously discover, build, and activate publisher first-party audience data without exposing or transferring the underlying data itself. The announcement, dated March 12, 2026, positions the integration as an early live demonstration of the Advertising Context Protocol (AdCP) operating across real programmatic infrastructure.
The deal connects two companies that have been collaborating on agentic advertising standards since October 2025. Optable, a Montreal-based data collaboration and identity platform founded in 2020 and employing between 11 and 50 people, was a founding member of AdCP when it launched alongside Scope3, Yahoo, PubMatic, Swivel, and Triton Digital on October 15, 2025. PubMatic, listed on the Nasdaq under the ticker PUBM, introduced AgenticOS on January 5, 2026, positioning it as the first operating system built specifically for autonomous advertising execution across premium digital environments.
What is new today is not the ambition - it is the specific technical implementation.
Buyers using PubMatic AgenticOS gain direct access to Optable's agent-driven audience planning capabilities. According to the announcement, advertisers can analyze publisher first-party data signals - including identity, contextual, and audience signals - and then activate custom data-targeted campaigns more efficiently across multiple publishers simultaneously. Critically, this happens without the underlying data leaving the publisher's environment. No raw data is exposed or transferred in the process.
Activation executes through PubMatic Activate, the company's direct-to-supply media bidder. This design choice is technically significant: publishers and their partners benefit from agentic audience workflows without needing to adopt new protocols or modify existing infrastructure. For sellers working with Optable, the Audience Agent surfaces high-value audiences to buyers on PubMatic automatically, tapping into what the companies describe as growing agentic demand from advertisers.
"This partnership ensures that publisher first-party data doesn't just inform planning - it creates a layer of audience intelligence that directly powers activation and programmatic buying decisions," said Vlad Stesin, CEO of Optable. "By integrating our Audience Agent into PubMatic AgenticOS, we are enabling buyers and sellers to collaborate securely, uncover new audience value, and activate it faster without compromising privacy, data or control."
The statement reflects a specific architectural argument: that first-party data should remain with the publisher while still powering the decisions made by buyer-side agents in real time. This stands in contrast to older data marketplace models where data was frequently transferred, aggregated off-platform, or made available in ways that reduced publisher control.
PubMatic AgenticOS, launched in January 2026, operates on NVIDIA-accelerated computing infrastructure embedded within PubMatic's global transaction systems. According to PubMatic's technical specifications, the infrastructure delivers sub-millisecond response times, up to five times faster decisioning, and reduced auction timeouts. The platform processes approximately 78 trillion impressions quarterly. Early testing with agentic campaigns demonstrated campaign setup time reductions of 87% and issue resolution improvements of 70%.
AgenticOS operates across three layers. The infrastructure layer handles large-scale, low-latency inferencing across millions of advertising transactions per second. The application layer integrates agentic capabilities that interpret intent expressed through protocols including AdCP and the Model Context Protocol. The transaction layer connects agentic decisioning directly to PubMatic Activate, supporting real-time transactions across Programmatic Guaranteed and Private Marketplace deals.
Kyle Dozeman, Chief Revenue Officer, Americas at PubMatic, framed the Optable integration within that architecture. "AgenticOS was built to connect intelligence to execution," Dozeman said. "Integrating Optable's audience capabilities allows AI agents to discover high-value, custom audiences and then move seamlessly to RTB activation across premium supply."
The phrase "seamlessly to RTB activation" is worth unpacking. In traditional programmatic workflows, audience discovery and activation are separate steps, often handled by different platforms with different data access agreements. The integration described today collapses that sequence: an agent discovers audiences through Optable's system, builds targeting parameters, and executes directly through PubMatic's real-time bidding infrastructure in a single, connected workflow.
The partnership also represents what both companies describe as an early example of the Advertising Context Protocolin action. AdCP, which launched on October 15, 2025 with six founding members including Optable and PubMatic, is an open-source technical standard built on Anthropic's Model Context Protocol. It defines how AI agents communicate campaign requirements, match inventory, and execute buys across platforms without requiring custom integration work for each system.
The significance of framing this Optable-PubMatic integration as an "early example" of AdCP in practice is considerable. Since the protocol's launch, industry commentary has been mixed. Industry veteran David Kohl warned in October 2025that AdCP represented premature infrastructure development before structural problems in programmatic advertising had been addressed. Ad tech analyst Ari Paparo questioned the protocol's media buying viability in November 2025, while supporting its creative automation specifications. The broader industry debate around agentic AI standards has persisted since the autumn of 2025, with notable holdouts: Google, Amazon, The Trade Desk, and Microsoft had not signed up for AdCP at its October launch.
Today's announcement does not resolve those debates. What it does provide is a concrete, production-level example of agent-to-agent collaboration - what the companies call a demonstration of how "privacy-safe audience intelligence can directly power real-time programmatic execution on the open web, without relying on closed platforms or centralized data systems."
That framing is itself a competitive positioning argument. The emphasis on the "open web" and the explicit contrast with "closed platforms" situates the partnership within broader programmatic market dynamics, where independent supply-side platforms have faced pressure from walled gardens operating proprietary data and buying systems.
Optable's platform is described as a data collaboration and identity infrastructure built for the advertising ecosystem. According to its company overview, the platform enables publishers, media owners, advertisers, and partners to collaborate on audience insights, campaign planning, activation, and measurement. Its LinkedIn page, verified in August 2024, lists the company's founding year as 2020 and its headquarters in Montreal, Canada.
The company's Agentic Collaboration framework is the conceptual backbone of the product: it transforms manual, human-mediated workflows into what it describes as intelligent agent-to-agent collaboration. Operationally, this means the Audience Agent can surface relevant publisher audiences to a buying agent on PubMatic's side without a human analyst in the loop at every step. The platform maintains what the company calls an identity graph built across multiple sources, enriched with partner datasets, and capable of activation across channels.
Optable's history with PubMatic and the broader agentic advertising ecosystem is longer than this particular announcement. Beyond co-founding AdCP, Optable was listed as a participating organization in the IAB Tech Lab's Container Project Working Group, which developed the Agentic RTB Framework released for public comment in November 2025. More recently, the IAB Tech Lab expanded its Agent Registry on March 11, 2026 - one day before this announcement - adding Optable among five new entrants. The Optable Audience Agent carries a "Private" deployment classification in the registry, the only such entry among the 10 listed agents, reflecting the company's focus on privacy-preserving audience infrastructure where data remains within controlled environments.
The partnership's stated benefits differ depending on which side of the transaction is considered. For publishers, the integration is designed to increase addressability and improve yield without adding operational complexity. According to the announcement, sellers working with Optable do not need to adopt new protocols or modify existing infrastructure to participate in agentic audience workflows. The activation path runs directly through PubMatic Activate, the same infrastructure publishers already use.
For advertisers, the integration provides more direct access to premium audience signals across PubMatic's marketplace. The promise is more precise audience analysis, faster campaign execution, and better outcomes - though neither company provides specific performance metrics in the announcement to substantiate those outcomes at this stage.
PubMatic's Activate platform has shown meaningful commercial traction in recent quarters. Buying activity on Activate more than doubled sequentially in the second quarter of 2025 compared to the first quarter, according to PubMatic's Q2 2025 earnings disclosure. That growth trajectory provides context for why adding a specialized audience intelligence layer to Activate - through Optable's Audience Agent - is a logical extension of the platform's existing commercial momentum. PubMatic also reported CTV revenue growth exceeding 50% year-over-year during the same period, reflecting broader programmatic advertising demand for premium inventory.
PubMatic's more recent financial results, released in February 2026, framed the company's agentic investment as central to its growth narrative. PubMatic CEO Rajeev Goel stated that he expected 25% of all digital advertising to execute autonomously via agentic AI by 2028, rising to 50% by 2030. Those projections, ambitious as they are, set the strategic context within which today's Optable partnership is positioned.
The identity challenge underpinning this partnership reflects a well-documented industry condition. Shared identifiers and mobile IDs have been declining in reliability as privacy regulations tighten and browser and operating system restrictions narrow tracking capabilities. According to Optable's platform documentation, the company helps media owners, publishers, and platforms unlock the value of their first-party data precisely because third-party signals have become less dependable.
Triton Digital and Optable previously partnered in February 2025 on an audio advertising identity solution, demonstrating 2.4 times higher ad fill rates through a system that enabled publishers to leverage user insights while maintaining privacy compliance. That implementation provided Optable with operational experience building server-to-server identity infrastructure across large media organizations, including iHeartMedia and Bell Media.
The current PubMatic integration extends that identity-first architecture into a real-time bidding environment. Rather than the data clean room model - where audiences are analyzed in an isolated environment and results delivered in aggregate - the Optable-PubMatic approach allows the Audience Agent to operate directly within the agentic buying workflow, surfacing signals without exposing underlying personal data to the buyer's systems.
The agentic advertising category is maturing rapidly, but several practical challenges remain unresolved. Industry analysis published in January 2026 documented consistent reluctance among advertising platforms and agencies to grant AI agents autonomous control over bidding decisions. Major platforms including Yahoo DSP, PubMatic, Amazon Ads, and Google Ads all implemented agentic capabilities while maintaining human oversight at the point of spend.
Whether the Optable-PubMatic integration changes that dynamic in practice - or whether the "autonomous" audience discovery still requires human approval before activation - is not fully specified in the announcement. The companies describe a workflow where buyers can "autonomously discover, build, and activate" audiences, but the operational details of approval gates and human oversight remain to be seen as deployments scale.
Who: Optable, a Montreal-based data collaboration and identity platform founded in 2020 with 11-50 employees, and PubMatic (Nasdaq: PUBM), a supply-side advertising technology company founded in 2006 and processing approximately 78 trillion impressions quarterly.
What: Optable's Audience Agent has been integrated into PubMatic's AgenticOS, enabling buyer-side AI agents to autonomously discover, build, and activate publisher first-party audience data signals - including identity, contextual, and audience signals - without exposing or transferring the underlying data. Activation executes through PubMatic Activate, the company's direct-to-supply media bidder. The companies describe the integration as an early live example of the Advertising Context Protocol operating across production programmatic infrastructure.
When: The partnership was announced on March 12, 2026. Background context spans from October 15, 2025, when both companies co-founded the Ad Context Protocol, through January 5, 2026, when PubMatic launched AgenticOS, to March 11, 2026, when Optable joined the IAB Tech Lab Agent Registry.
Where: The integration operates within PubMatic's global infrastructure, headquartered in Redwood City, California, and spans premium publisher supply across PubMatic's marketplace. Optable is headquartered in Montreal, Canada.
Why: Publisher first-party data has become increasingly important as shared identifiers and mobile IDs have declined in reliability due to privacy regulation and platform restrictions. The integration aims to make publisher first-party data actionable within agentic media buying workflows without requiring data transfers or infrastructure changes from publishers. Broader industry context includes a competitive dynamic between open-web programmatic platforms and closed, walled-garden advertising ecosystems, with both companies arguing that open interoperability through standards like AdCP offers structural advantages for independent publishers and buyers.