The deployment of AI Overviews by Google in France since July 22, 2026, is reshaping organic visibility. Pages that merely target a keyword without structuring a direct answer are losing impressions to the sources cited in AI summaries. This context requires a rethink of content production, technical linking, and authority signals well beyond traditional SEO practices.
Generative Engine Optimization: structuring content for Google’s AI answers
AI Overviews do not pick sources at random. We observe that the pages cited as sources in these summaries share specific characteristics: a response formulated within the first hundred words, sourced numerical data, and a structure that facilitates passage extraction. The term that emerges to describe this discipline is Generative Engine Optimization (GEO).
In practical terms, GEO complements SEO without replacing it. Well-positioned content in traditional results may never appear in an AI Overview if it lacks authority signals or if its answers are buried in a 200-word paragraph. Conversely, a mid-ranking page with a clear structure may be selected as a source.
To delve deeper into this approach beyond just AI Overviews, the SEO advice from Agence Web Marketing also covers the technical fundamentals that remain the foundation of any sustainable visibility strategy.
We recommend treating each page as a unit of response. If an article covers three distinct questions, it should have three identifiable blocks, each with its H2 containing the question and a first paragraph that formulates the answer. The rest of the block develops, nuances, and illustrates.
Transactional and local queries: why local SEO is changing with AI Overviews
AI summaries now extend to transactional and local queries. This changes the game for service and local businesses that relied on the local pack or traditional organic results to capture clicks.
When Google generates an AI response for a query like “emergency plumber + city,” the visibility of the Google Business Profile alone is no longer sufficient. The website page must also be structured to be cited: explicit mention of the geographical area, factual description of the service, pricing or range if available.
Adapting service pages to an extractable format
A local service page that consists of generic commercial text will not be retained by the AI engine. We find that high-performing local GEO pages incorporate:
- An opening paragraph that addresses the transactional intent in less than three sentences (what, where, within what timeframe)
- Correctly implemented LocalBusiness or Service structured data, which provides the engine with actionable signals
- Content that clearly distinguishes each service offered instead of grouping everything on a catch-all page
This granularity by service and area creates more entry points for AI Overviews.
Authority signals and editorial freshness: the criteria for selecting AI sources
Google does not cite just any page in its summaries. Editorial freshness and authority signals weigh more heavily than before in the selection of sources. Content dated from 2022, even if well-positioned, is unlikely to be retained against an article updated in 2026 on the same topic.
Freshness is not just about changing the publication date. We recommend adding recent verifiable information: mention of a regulatory update, integration of a sourced industry figure, or addition of a section that addresses an observed evolution in the field.
Building thematic authority on a cluster of pages
A site that publishes only one article on a topic will not be perceived as a reliable source by the AI engine. Thematic authority is built through a cluster of interconnected content: a pillar page that covers the main topic, linked to satellite pages that address each sub-theme in depth.
The internal linking between these pages must be logical and bidirectional. Each internal link conveys contextual relevance, which enhances the cluster’s ability to appear as a source in AI Overviews for multiple query variations.
Measuring SEO visibility in the context of AI Overviews: indicators to track
Classic metrics (average position, click-through rate) are no longer sufficient to evaluate organic performance. The decline in clicks related to AI responses necessitates tracking new indicators. Several analyses from summer 2026 document a general decrease in the click-through rate as soon as an AI Overview appears for a query.
We recommend cross-referencing three types of data:
- Tracking queries triggering an AI Overview in your sector, via Search Console and third-party tools that detect the presence of these blocks
- The rate of your domain appearing as a cited source in AI Overviews, distinct from simple organic positioning
- The evolution of qualified traffic per page, as a page cited as an AI source may generate fewer clicks but a higher conversion rate
- Impressions without clicks, which reveal queries where the AI response satisfies the user without visiting your site
This tracking allows for arbitration between content to optimize for GEO and those that perform better through traditional SEO.
SEO in 2026 does not oppose traditional referencing and optimization for AI engines. The two reinforce each other when the site’s structure, editorial quality, and authority signals are aligned. The difference lies in the ability to produce content that meets a specific intent in a format that the engine can extract and cite.
