Ogury Launches SONA, an Agentic-AI Targeting Engine

Ogury has launched SONA, a new solution built on agentic AI that aims to improve consistency between how advertisers plan their audiences and how those audiences are actually activated across live campaigns. The company frames SONA as an attempt to close a persistent gap in programmatic execution: the divergence between the audience an advertiser defines at the planning stage and the impressions that audience strategy ultimately delivers at scale.

The planning-to-activation gap

In most programmatic workflows, audience strategy and campaign activation are handled as distinct phases. Planners define cohorts, contextual segments, or behavioral profiles, and those definitions are then translated into targeting parameters inside a DSP and matched against available supply. The translation is rarely lossless. Segment availability, bid-stream signal quality, supply fragmentation, and frequency dynamics all introduce drift between the intended audience and the audience that is actually reached.

SONA positions itself against exactly this problem. Rather than treating planning and activation as a handoff, Ogury describes the product as continuously reconciling the two — using agentic AI to keep the activated audience aligned with the planned audience as a campaign runs. The promise is that what was modeled in the planning phase remains the operative target during delivery, instead of degrading as the campaign optimizes toward whatever performance signals are most readily available.

What "agentic AI" implies here

The term agentic AI refers to systems that don't just generate recommendations but can take iterative, goal-directed actions toward an objective with limited human intervention. Applied to media activation, that suggests SONA is intended to monitor in-flight delivery, detect when the realized audience diverges from the planned definition, and adjust targeting or bidding behavior to pull delivery back in line — repeating that loop across the campaign lifecycle.

For ad ops teams, the relevant question is where this sits in the stack. Audience consistency tooling is only as good as the signals it can act on and the levers it can pull. If SONA operates upstream of activation — shaping segments and targeting logic that are then executed through standard programmatic pipelines — its impact depends on how cleanly those definitions survive the OpenRTB bid stream and DSP-side optimization. If it has direct control over activation, the consistency story becomes more credible but also raises the usual questions about transparency, measurement, and how "alignment" is quantified.

Why audience consistency matters for programmatic

Audience drift is not a cosmetic concern. When activated delivery diverges from the planned strategy, advertisers pay for reach that doesn't match the intent behind a campaign, and measurement against the original objective becomes unreliable. This is particularly acute in a post-third-party-cookie environment, where deterministic identity signals are scarcer and platforms increasingly lean on modeled audiences, contextual approaches, and probabilistic matching. Each of those introduces additional uncertainty between plan and execution.

Ogury has historically built its identity around what it calls personified advertising — audience targeting that does not rely on tracking individual users. A tool that reinforces consistency between planning and activation fits that positioning: if the audience is defined through modeled or contextual signals rather than user-level identifiers, maintaining fidelity between plan and delivery becomes both harder and more valuable. Consistency tooling is, in effect, a way to make privacy-conscious targeting performant enough to compete with identity-based approaches.

The competitive context

SONA arrives amid a broad wave of agentic-AI announcements across ad tech, where vendors are racing to attach autonomous decisioning to media planning, creative, and activation. The differentiation will come down to whether these systems demonstrably reduce waste and improve outcomes against an advertiser-defined audience, or whether "agentic" remains a marketing layer over existing optimization. For buyers, the practical test is measurable: does activated delivery match the planned audience more closely with SONA in the loop than without it, and can that be independently verified?

For publishers and supply-side stakeholders, tools that tighten audience targeting on the buy side have downstream effects on which inventory gets bid on and at what density. More precise activation can concentrate demand on supply that matches advertiser intent, with implications for fill and yield across the open programmatic marketplace. As agentic activation tooling matures, the interplay between buy-side autonomy and sell-side transparency will be worth watching closely.

Stay on top of video ad serving and programmatic. Follow Adelerate.