Social Rocket
Scope3 is becoming Apostra and is aligning its business much more strongly with agentic media buying. The new presentation no longer positions the platform primarily around sustainability, but as infrastructure through which AI agents are meant to coordinate ad planning, purchasing and optimisation across multiple channels.
The change is more than a name change. It shows how part of the adtech industry is moving from classic programmatic automation towards agent-based media planning.
Apostra does not want to automate individual impressions alone
On the official Apostra website, the company describes a model in which humans set goals, rules and brand guardrails, and agents then take over the operational work: finding opportunities, comparing prices, requesting proposals, allocating budgets, booking campaigns and monitoring delivery.
This is not meant to cover only display or classic programmatic channels. Apostra names, among others, CTV, audio, out-of-home, social, sponsorship and experiences.
The central difference from previous programmatic logic lies in focus. Instead of auctioning every single impression in milliseconds, agents are meant to work more at the level of campaigns, portfolios and negotiations.
AdCP provides the shared language between agents
Technically, this approach is built on the Ad Context Protocol (AdCP). The open standard was launched 2025 by a consortium including Scope3, Yahoo, PubMatic, Optable, Swivel and Triton Digital.
AgenticAdvertising.org describes AdCP as a shared language for AI agents in the advertising ecosystem. The standard is meant to make discovery, planning, buying, creative processes and reporting interoperable across different systems.
This matters for marketers because the real hurdle for AI agents in media buying is not only the quality of individual models. Each platform has its own APIs, taxonomies, approval processes and reporting logic. A common protocol is meant to prevent individual integrations having to be built again for every channel.
Agentic media buying is already broader than a single provider
Apostra is therefore not alone. Other adtech providers are also building specialised agents for research, planning and campaign activation. Srocket has already analysed how DeepIntent combines several AI agents with Cora for pharma media planning.
In principle, what we described in our overview of AI in media buying still applies: more automation makes independent control of budget logic, recommendations and measurement more important, not less.
The current Google adtech overhaul also shows how strongly interoperability and access to advertising infrastructure have become strategic questions. Agentic systems shift this discussion further onto the level of protocols and automated negotiations.
What could actually change for media teams
For media teams, the most exciting question is not whether an agent can optimise individual clicks faster. Today's platforms already do that. The bigger change would come if agents could actually understand briefs across channels, compare inventory, negotiate terms and shift budgets.
This would move part of the work from manual platform operation towards rule-setting, data quality, approval and control. That is precisely why the AdCP approach also emphasises human authorisation and governance.
This development is still early. Standards and business models are changing quickly, and an open protocol guarantees neither better prices nor better campaigns. For advertisers, Apostra's rebranding is nonetheless a clear signal: agentic advertising is evolving from individual AI features into its own infrastructure layer in media buying.