This led marketing specialists to order agencies like Havas, Dentsu and Kepler to revise their organic research approach – 57% of marketing specialists have changed their research strategies since IA previews were launched in 2024, according to a survey by Agency NP Digital – while leaving the paid media part of their intact research strategy. This status quo will not last long.

Google increases its deployment of IA previews, the functionality now appearing on 47% of search results pages, according to requestphere. Most of these represent the types of research queries that brands do not often submit against, large research requests that imply that a web user is located higher in the sales funnel, but media buyers suspect that they will not be limited for a long time.

Meanwhile, the use of chatgpt and perplexity for research continues to climb. The Active Chatgpt user base reached 400 million at the start of this year while Perplexity added 2 million active users to reach 22 million in total between October and the first half of 2025, per application company. Perplexity is currently playing with its advertising product, while Chatgpt is expected to launch its own in the near future.

Luminous dashboard

With these silhouettes on the horizon, media buyers closely monitor a range of indicators.

“I have never watched the reference traffic in my life so much,” said Eric Hoover, director of SEO at Kepler’s. He noted that the references have taken on disproportionate importance as a key indicator that web visitors arrive from a summary generated by AI.

The volume, the prominence and the tone of the quotes in the summaries of the AI ​​are also worth looking at. “The new classification signal is the number of quotes … It is a really strong signal of a good brand presence,” said Hoover.

Research practitioners also closely monitor the cost rates by CLIC (CPC) of the Google research announcement inventory. CPCs increased by 9% in the second quarter of the year, according to Tinuiti. The metric reflects the modification of the dynamics of supply and demand – some of which are due to the behavior of AI research users, and others to tinker Google with the landscape of search results pages.

Since the IA previews appear at the top of the results pages, they pushed the paid announcements that appear next to them further down the search results. Marketing specialists have prioritized spending on keywords less likely to generate an AI summary, which raises CPCs. “Everyone is continuing the same requests now, and it becomes really competitive,” said Jeff Eisenfeld, research director at Media by Mother.

“We note that the AI ​​glimps re-entertain the research behavior and, in many cases, leading the CPC while advertisers compete for fewer clicks,” said Brooke Hess, VP paid Media at NP Digital, in an email.

But AI’s overviews are not the only culprit. Today’s serp is a crowded window panel rather than a blue link repertoire; Visual search results and suggested research features are also blamed. “It is difficult to determine the attribution,” said Daniel Toplitt, EVP, Research and Digital Experience in the KINESSO performance marketing unit of IPG.

And Amazon’s recent steep withdrawal of the investment in Google’s purchase advertisements has further complicated the situation. Price fluctuations can be the consequences of the retirement of the electronic commerce giant, or evidence that brands in general modify their expenses in response to AI research.

“The working thesis is that the decline in organic and paid research clicks, invited to a behavior increased by zero clicks, exerts ascending pressure on CPCs,” added Hess. “That said, other factors such as the increase in advertisers’ competition, seasonal trends or strategic offers could also play significant roles.”

Short -term expenses and long -term implications

For the most part, the overall expenses of paid research remained stable. “If [they] See the opportunity to capture more income, so they will be ready to spend more, of course. [But] I don’t think a customer came to use [for an] Increase due to AI previews, “said Kenneth Yau, paid research director at Dentsu. But a short -term manual appeared for brands that have brought a blow of web traffic – often those in the electronic commerce category. They increase paid expenses for the brand’s keywords and not mark as a way to “protect” the research that matters for sales conversions.

“We sometimes recommend increasing short-term remunerated media investments to protect visibility, because the research led by AI limits organic scope,” said Rachel Klein, vice-president of owners and won at Wpromote. But it is a short -term pivot, a “dressing” which could be less effective should (well, when) the IA previews begin to appear on increasingly searching results.

“The long-term success in AI research (and research in general) depends on a holistic strategy that requires an emphasis on construction authority, and not to exchange organic tactics to pay,” she warned. Brands cannot ignore the “fundamental” work required to take up the zero challenge for a long time.

This fundamental work may well require rethinking what, from the point of view of an advertiser, research is really.

“Zero click research affects research. This affects public relations, this affects trade,” noted Michael Sondak, please, and the research manager of Omnicom Media Group North America.

As such, a zero -click -free response may require a response joining the points of each of these quarters – which reflects the overall strategic objectives of a brand, and not only the need to hunt web users further in the sales funnel.

“I am less concerned with an increase in CPC. The question I ask my teams from the remunerated media point of view is: do you have challenges to strike the commercial objectives that we have registered to deliver? ” said Sonnedak. “If we have registered in Roas [or] The slopes, do we deliver that? It is the most important metric for me.