Google's AI Overviews arriving in France are not just another search option. They change an implicit contract that has structured the web for years: sites publish, Google indexes, users click. When a synthetic answer appears directly above links, the click becomes less automatic.
For users, the appeal is clear. A long query can receive a structured answer with links and possible follow-ups. For publishers, the situation is more ambiguous. Being cited in an AI answer can reinforce a brand's authority, but it does not guarantee a visit. Not being cited can make content invisible even if it still ranks well in classic results.
SEO no longer stops at rank
SEO used to follow a relatively readable logic: publish a useful page, make it understandable to search engines, earn links and improve its position. AI Overviews add another layer: the engine can extract part of the answer, recombine the context and display a summary before the user opens a site.
Ranking still matters, but it is no longer the only signal. A page can rank well and lose traffic if the AI answer satisfies the main intent. Conversely, a brand can gain visibility if it is cited in the summary, even when the visit does not always follow.
What publishers need to measure
Organic traffic remains the first metric, but it should be combined with other signals: queries where an AI answer appears, brand presence inside those answers, click-through rate when cited, pages most affected and changes in conversions from search.
This measurement will be imperfect at first, because analytics tools do not always show what happened inside the result page. Teams will need to combine Search Console, rank tracking, manual tests, query panels and observation of pages that lose traffic without losing classic rankings.
Make content more citable
Fragile content lacks clear signals: no date, no author, few sources, vague titles, long undifferentiated paragraphs, marketing claims and no direct answer. In a generative environment, those weaknesses become more visible.
Stronger content answers a precise question, shows sources, separates facts from analysis, dates updates and gives the reader a reason to click beyond the summary. That reason can be methodology, comparison, proprietary data, field expertise, a demonstration or an archive better structured than the engine's short answer.
Value moves toward trust
AI Overviews do not remove the need for sources. They make it more explicit. A user may read the summary to understand a topic quickly, then click to verify, go deeper or get an identifiable point of view. Recognizable editorial brands therefore become more important, not less.
This requires a product strategy. Media outlets, brand blogs and specialist sites need to build direct reader relationships: newsletters, RSS feeds, internal search, social accounts, communities, alerts and updated dossiers. Search traffic does not disappear, but it becomes less predictable.
A chance for specialist sites
The battle will not be only about volume. Sites that publish verifiable, well-structured and genuinely useful content can remain visible, especially on queries where users want to compare, decide or understand the limits of an automatic answer.
AI Overviews do not kill web search. They move value toward being a reliable, recognizable and useful enough source that users still have a reason to leave the results page.
