AI Overviews are coming: an advertiser's homework before it hits France

Google is about to roll out AI Overviews in France, a feature already live for months in markets such as the U.S., U.K. and Germany. Displayed above traditional results, these AI-generated answers synthesize multiple sources to respond directly to certain queries. They're expected to mainly affect informational searches, a share of which can now be satisfied without a single click.

Early data suggests the impact will vary widely. In Germany, SEO software firm Sistrix estimates that the top organic position's click-through rate drops from 27% to 11% when an AI Overview is shown. But averaged across all queries studied, the overall decline is 6.6%, with sharp disparities across sectors. In other words, not every brand will be affected the same way.

To help advertisers prepare, we spoke with several search and AI-visibility specialists. Here are the four priority workstreams to get moving on over the coming weeks.

The first mistake would be assuming AI Overviews make SEO obsolete. "It's obviously wrong to think SEO is dead, replaced by GEO or AEO," warns Vincent Cézard, digital trading lead at agency Biggie. Google's generated answers still rely on pages the engine is able to discover, crawl and properly evaluate.

In other words, the fundamentals remain table stakes. Havas Market France deputy managing director Marina Houdayer recommends starting by checking "site indexability, internal linking, technical performance, structured data and content readability." Google itself notes that no AI-Overviews-specific markup is required: a page above all needs to be properly indexed and follow standard Search best practices.

Beyond these technical aspects, Vincent Cézard points to the importance of E-E-A-T (Experience, Expertise, Authoritativeness and Trustworthiness), still one of Google's main quality frameworks. "The goal is to position yourself as a reference, to bring real added value," he explains. That means publishing reliable information, verified data, expert opinions, or first-hand experience that's hard to replicate.

That said, good rankings no longer guarantee an appearance in an AI Overview. Nearly 30% of domains cited in AI Overviews don't even appear on the first page of the associated organic results, according to an academic study covering more than 55,000 queries.

Ranking well is still an advantage, but it's no longer sufficient to be picked as a source. The real difference will come down to brands' ability to answer the queries where Google chooses to show an AI Overview... and to publish content useful and distinctive enough to be woven into the answer.

Not every Google search will be affected by AI Overviews the same way. Before overhauling their editorial strategy, advertisers should first identify the queries where Google is most likely to answer directly in place of their site.

Early rollouts show that AI Overviews appear mostly on informational, comparison-based or question-phrased searches. They showed up on 13.7% of searches analyzed overall, but on 64.7% of queries phrased as a question, according to an academic study covering more than 55,000 queries.

For Marina Houdayer, this shift forces brands to move beyond traditional keyword logic. "Content needs to be built around search intent rather than keywords alone. Conversational and informational queries are becoming increasingly important."

To spot these opportunities – or risks – Houdayer recommends widening the sources feeding the editorial strategy: "Search Ads data, internal searches, customer service verbatims, questions raised by sales teams, or customer reviews."

The question is no longer just which keywords a brand ranks for, but which situations it wants to be visible in.

Once that mapping is done, the next step is deciding which content deserves reinforcement. Vincent Cézard recommends identifying "which pages still answer a real intent, which ones repeat information available everywhere, and which ones bring data, expertise or an experience unique to the brand."

He advises structuring content around precise questions and quickly identifiable answers. Without turning every page into an FAQ, it becomes essential to get to the point fast and make it easy for Google to extract the information.

"By producing answers that are too easy to summarize, brands can end up giving people fewer reasons to visit their site," he warns. The goal is to make information usable by Google while keeping enough added value to justify the click: comparisons, proprietary data, methodologies, calculators or first-hand accounts.

Cézard recommends scaling back purely definitional content, which AI Overviews absorb easily, in favor of buying guides, comparisons and experience-based content.

Beyond format, substance becomes a differentiator. The more a piece of content relies on generic information available on dozens of sites, the more easily it risks being replaced by an AI-generated summary.

Brands therefore have an interest in publishing content their competitors can't easily replicate: their product knowledge, their data, customer feedback, their tests, their comparisons or their expert advice. It's precisely this distinctiveness that increases their chances of being cited by AI Overviews... while giving users a real reason to keep browsing. For Frédérique Goubert, head of search at iProspect, this shift could reshuffle the deck. "A specialized brand, able to provide a unique answer on a specific topic, can win out over a much bigger but less distinctive player," she says.

Optimizing your own content won't be enough. Generative engines constantly cross-reference information from multiple sources to build their answers. A brand's reputation is therefore shaped as much off-site as on its own pages.

A comparison published by a specialist media outlet, a strong rating on a review platform, or a detailed YouTube video can today carry as much weight – sometimes more – than a perfectly optimized page on the brand's own site.

Goubert recommends "auditing how the brand is described on comparison sites, review platforms, media outlets, specialist sites and communities." A claim made on the official site loses credibility if it's contradicted elsewhere.

Vincent Cézard sums up this shift as a kind of "link-building 2.0." "The goal is no longer just to earn inbound links, but to be regularly and credibly associated with the topics the brand wants to be seen as an authority on." A product test, a comparison, or an analysis in a specialist outlet often carries more context than a plain backlink.

Third-party sites account for 85.7% of the citations used by various models, against 14.3% for brands' own properties, according to a study of more than 167,000 citations. While this study isn't specific to AI Overviews, it illustrates how much weight external sources carry in building a company's reputation.

That doesn't mean artificially multiplying mentions, though. Google continues to state that its systems favor quality content and push back on spam tactics. The real work lies in identifying the sources that genuinely structure each market – specialist media, comparison sites, creators, review platforms or communities – and then building a credible presence there.

Within this ecosystem, YouTube stands apart. It is the most-cited platform, ahead of Reddit, according to an analysis of more than a billion citations from various AI assistants.

For Frédérique Goubert, "the tightening links between Search, Gemini and YouTube should further reinforce its role in discovery journeys." Brands therefore have an interest in polishing both the substance and the form of their videos: titles, descriptions, chapters, subtitles and transcripts all become comprehension signals for Google.

The arrival of AI Overviews doesn't just challenge SEO playbooks. It also shakes up how performance is measured. A drop in traffic doesn't necessarily translate into a drop in business, while a brand can influence a purchase decision without generating a single click.

"The main lesson from countries where AI Overviews is already live is that you can no longer think purely in terms of clicks," Frédérique Goubert says. Vincent Cézard also warns against "the temptation to systematically offset any drop in organic traffic with more paid media spend."

For Marina Houdayer, the best way to avoid that trap is to "keep indicators directly tied to the business: CPA, ROAS, incrementality or sales contribution."

Traditional SEO metrics remain relevant, but they need to be gradually complemented with new signals: how often a brand is cited in generated answers, share of voice against competitors, branded-search trends, or direct traffic.

Visibility itself becomes harder to interpret. A brand can be recommended within an answer without being explicitly cited as a source. Conversely, it can be used as a reference without gaining much exposure.

In one case, it directly influences the user's choice; in the other, it mostly reinforces its credibility with the model. These two forms of visibility don't carry the same value and will need to be tracked differently.

This shift doesn't just change dashboards. It also challenges how teams are organized. Product-page quality is often owned by e-commerce teams, customer reviews by CRM, expert commentary by communications, videos by social, and Search investment by media teams. Yet Google now pulls all of these signals together to build its answers.

Marina Houdayer accordingly recommends "setting up cross-functional governance" able to identify priority queries, regularly audit generated answers, and arbitrate actions across content, technical, product, PR, social and media.

At its core, AI Overviews blur the lines between SEO, content, communications, CRM, social and media. Organizations that keep steering these levers separately risk ending up with a fragmented view of their visibility, even as Google now evaluates them as a coherent whole.

The arrival of AI Overviews, then, doesn't mark the end of SEO – or even of search. What it does is reshuffle the criteria for visibility. Brands that can produce distinctive content, build authority beyond their own site, and measure their presence by more than traffic alone will be best placed to make the most of this new stage of search.