Search AI: the new visibility battle is already here
Reading time
7
min.
For twenty years, online research has been based on a relatively stable balance: keywords, blue links, traffic returned to sites. This balance is in the process of breaking. The rise of generative AI is changing both user behaviors and the historical hegemony of GAFAM. This break is leading brands to urgently reassess their visibility strategies: not only where they appear, but above all how they are understood, cited and relayed by new research agents.
The upheaval of uses
The “zero click” phenomenon, when the answer no longer leads to the site: One number alone summarizes the transition under way: 58% of Google searches in the United States do not result in any clicks. For 6 out of 10 searches, the user gets the answer directly in the engine interface. The engine no longer returns traffic: it becomes the only place of response. This phenomenon marks the end of a paradigm: visibility is no longer measured only by traffic, but by the ability to be integrated into the responses generated by AI.
Longer, more contextual queries: While a traditional query relies out of 4 keywords on average, the prompts sent to the AIs contain 23 words on average. This semantic shift reflects a profound change: we are no longer looking for pages, but structured responses, contextualized, sometimes argued. Users no longer “scan”, they delegate.
New players in global research: Traditional research still represents 90% of uses, but the growth momentum is clearly on the side of AI engines. A conviction: it is their logic, and not that of historical SEO, that will structure the future of visibility.
In addition, Research is no longer limited to engines, it becomes social: it extends to conversational and video platforms.
- YouTube is the 2nd search engine in the world,
- 68% of Gen Z use it to search,
- 82% of French Gen Z is active on TikTok,
- 64% use TikTok as a search engine.
This shift towards short videos and UGC content is based on a simple logic: trust, speed, proximity. The LLMs also favor these sources: in June 2025, Reddit is more cited than Google in AI responses (40.1% of quotations versus 23.5%).
The big decoupling: the impression goes up, the click falls. The analyses show a clear dissociation between two indicators that have long been correlated:
- increased printing rate,
- drop in the click rate, sometimes — 35% on informational requests.
AIs increase consultation (scraping, analysis), but capture a large part of human traffic. This does not mean a loss of performance: on the contrary, on transactional sites, incoming traffic is less voluminous but more qualified, with observed examples ranging from +35% in revenue despite -20% in traffic.
This change makes it essential to monitor an indicator that has long been overlooked: The printing rate (understand here: in Search AI engines).
AI Overviews: when Google absorbs the response The arrival of AI Overviews — this paragraph generated on the fly, in response to keywords, in the search engine and before the traditional results (already visible in a large number of countries, imminent in France) leads to a significant drop in organic CTR:
- The CTR for the first result changes from 7.3% to 2.6% when an AI Overview is positioned at the top.
- A drop in -61% in the presence of this AI insert.
The SERP is no longer a crossing point: it becomes the final destination.
How can we remain visible in this new environment?
Owned Media vs Earned Media: a balance to be rethought. Today, AIs rely on:
- 10% off “owned” content (sites, blogs, internal resources),
- 90% on external sources (press, forums, UGC platforms, reviews, third-party publications).
Brands can therefore no longer rely solely on their website, even if it is optimized. Visibility depends on the ability to exist outside of its own platforms, in an ecosystem of multiple sources.
What AI engines prefer: to be quoted before being clicked. Two axes structure an adapted strategy:
1. Make your site a recognized source
The engines cite more:
- Contents recently updated: 66.7% of citations are for content < 1 year,
- sites with strong authority,
- structured pages (schema, FAQ, sources, BLUF...).
2. Broadcast your content beyond your site
Visibility today depends on:
- press articles,
- publications on community platforms,
- reviews, forums, UGC content,
- of the presence on social networks.
In particular, Reddit, YouTube, and TikTok are massively used as sources by AI agents.
Producing, developing, distributing: the obligation for brands. To exist in Search AI, brands need to set up a new editorial routine:
- Regularly produce hot, expert, structured content.
- Declare the same content in several formats : article, carousel, short video, podcast, practical sheet.
- Encourage UGC : opinions, feedback, conversations.
- Be present on several channels, because AIs intersect and prioritize sources.
Being “the owner” of your content is no longer enough. You have to be visible where AIs look for authority signals.
Measure differently: new KPIs to be integrated. Historical metrics - visitors, CTR, positions - become Insufficient. The Search AI approach requires the integration of indicators such as:
- number of mentions,
- Share of Voice IA,
- Feeling of the responses generated,
- AI readiness content,
- distribution of quotations by engine.
This data makes it possible to understand how the brand appears, is interpreted, and is cited in AI responses. New tools are coming onto the market, allowing us to go further than traditional SEO measures.
Conclusion: visibility depends more on one channel, but on an ecosystem
We are entering a new era where visibility is no longer just about being found, but about being Restrained and quoted by agents capable of synthesizing information even before a click occurs.
Brands must now think of their presence strategy as a network: a coherent set of content, signals, external sources and complementary formats.
The question is no longer “how do you attract traffic?” , but how to become a reference that AI engines will benefit from citing?