Media agencies use AI search insights to predict what audiences want

The methodology that Jen Cornwell, senior director of AI innovation at Tinuiti, and her team is developing marks the first steps toward building an audience graph for AI search, and is aiding both organic content and paid media strategies for clients, particularly regarding retail media spending. Since Tinuiti began using Profound in this way six months ago, it’s run 100 search audits on 10 clients, she said, but declined to name the clients.

“It’s another input. It’s another understanding of how people are consuming and starting to behave,” said Cornwell. While not the sole signal behind any given investment, it’s a valuable addition to the buyers’ toolbox when consumer search and online consumption habits are shifting into new patterns.

Earlier this year, agency businesses began offering their own zero-click search analysis tools, using their findings about how LLMs represented clients and product categories to fine-tune organic search and web strategies.

Tinuiti isn’t the only business working to tether AI search insights to practical outcomes.

Publishers such as Forbes are already exploring the practice in earnest. It’s been working to create audience cohorts based on data from SEMrush and Similarweb, that can be utilized by its editorial teams. “We create cohorts, meaning we understand the types of searches and the types of people that resonate,” said chief innovation officer Nina Gould, speaking on stage at the Digiday Publishing Summit Europe in Lisbon, Portugal, on Oct. 29.

Elena MacGurn, svp of search at Digitas, said the agency was using zero-click insights “to inform media activations, content production, marketing.” The agency recently launched its own AI search analysis tool, Model Sight.

At other agencies, incorporating zero-click data still resembles a workaround rather than a dedicated workflow — for now. 

“It gives us a wealth of data that we obviously didn’t have access to before, and it helps us understand, with a lot more precision, how to plan everything  more proactively from paid media to organic optimizations,” said Amanda Zanardi, associate director, earned and owned strategy at Wpromote.

Oli Williams, head of digital performance at Mediaplus UK, said the agency will run a campaign with an unnamed client later this month, using a media plan informed by AI search insights for the first time this year. 

“It is something where we’re factoring heavily into 2026 planning with clients,” he added.

Sam Clarke, head of search at Crossmedia, said the agency was in the early stages of expanding its use of the practice. The company began by carrying out its own analyses of ChatGPT search results, but has recently explored using tools from PKI and SEMrush; he didn’t share specific findings.

“It’s definitely got legs, especially if we can supplement it with some of those third-party solutions,” said Clarke. “I think it’s made us smarter in terms of how we invest our media.”

Some media buyers are wary of using LLM search data unless it’s gleaned directly from source — that is, LLM developers like OpenAI — something that’s not currently available.

That’s because, ultimately, tools like Profound (or agency-made solutions) are making inferences about the inner workings of LLMs; market research work that records information about what users are prompting for is similarly removed.

“You’re talking about multi billion dollar investments that can’t simply be made based on an assumption from a third-party tool,” said one holding company media buyer, who exchanged candor for anonymity.

They added: “We’re using the data to help inform and help steer… but not solely basing media buying decisions on LLM data.”