The world of ad tech is starting off the new year the only way it knows how: with a new agentic solution.
On Monday, Optable launched a new planner agent to automate the ad planning process for publishers. The tool analyzes advertiser campaign briefs, generates audience segments to target and reports back on measurement and outcomes once the campaign is live.
Hurry up!
Historically, the process of analyzing an RFP has been “slow and manual,” especially when the publisher is creating custom audience segments, said Jessica Hogue, chief data officer for consumer media at Hearst.
An agentic planner is more discerning than the human eye alone, thanks to the quantity of data it sits on and the speed at which it can analyze that data. And the new Optable agent has helped remove “a handful of baton passes” in the process, said Hogue.
Hearst and Optable have worked together to organize and taxonomize engagement across Hearst’s websites to help target audiences for specific advertisers or campaign goals. It’s like having the ocean “distilled and recommended back to you,” said Andrew Dumas, Optable’s US managing director. The alternative, he said, can mean manually typing in keywords or just hoping impressions are going toward the right audiences.
If a publisher receives an RFP for audiences interested in buying a car, for example, the obvious (and historically standard) response is to type in keywords like “auto” and “electric vehicle” as a proxy for potential buyers. They’re “probably not going to type in 58 car brands” and a list of other attributes that might fit someone planning to buy a car, said Dumas.
The agentic solution has no problem packing all those keywords and metadata into its response. It looks at all of the available data and proposes a “much wider swath” of specific audiences, Dumas added. It might recognize that people who live in Connecticut and play golf are more targetable for Mercedes-Benz and focus on those attributes rather than car-specific searches.
The AI’s goal is to identify opportunities beyond the intuitive targeting strategies the managed service version would bring.
The evolution of language
Publishers can engage with the agent, which is built on Claude, through standard chatbot natural language prompts. They can also upload documents, like results from a previous campaign, and “converse” with the AI to refine its knowledge of the advertiser’s particular audience targeting or brand strategy.
The agent is built on open standards, including MCP, as well as new industry protocols like AdCP, both of which Dumas describes as a “new language” built for agents that allows buyers and sellers to communicate directly. An agent from a holding company can reach out directly to Hearst and ask specific questions, like how many auto intenders they have across their properties, he added.
The interoperability makes it easier to communicate with holdcos and agencies that are just beginning to integrate with agentic workflows, said Hogue.
“This is a way for us to be ready and leaning into that opportunity,” she said.
Identifying solutions
Cookies are (famously) going to live on. (Alongside the Twinkie and the cockroach, third-party cookies may somehow outlast humanity.) But they’ve been largely abandoned by the big online ad platforms, which have the scale of data and media not to need third-party cookies. For publishers to ensure that ads are getting in front of the right audiences, signal loss will continue to make the data-driven ad process more challenging, Hogue predicts.
But the evolution beyond cookies could also be considered “signal gain,” said Dumas. Optable offers 15 alternative IDs, including UID2s, Lotame’s Panorama ID and Criteo ID. On the sell side, “you need to be flexible” and “ensure a high metric between what the buyer wants and what the seller has,” he said.
Hearst is optimistic that Optable’s planner agent will analyze various signals “with a speed and depth” that’s only possible with an AI platform, said Hogue.
Agentic workflows, she added, are the best way to tackle some of the “inherent friction” that publishers face in delivering strong outcomes-based campaigns.
