When looking at the statistics of a site in October 2026, one line stands out: the traffic referred by ChatGPT or Google AI Overviews appears in the reports, but the actual clicks to the source pages stagnate or decline. The innovations on the web are no longer just about new features or graphic redesigns. They now affect how content is consumed, cited, sometimes scraped, without the publisher deriving a measurable benefit.
AI Visibility and Real Traffic: What the New Analytics Reports Don’t Show
We’ve all experienced the situation: an article appears in a response generated by a conversational assistant, the site’s name is mentioned, but the analytics dashboard remains flat. AI search reduces clicks to source sites, and visits from assistants do not compensate for the losses.
The problem lies in the very mechanics of these tools. ChatGPT Search has indexed hundreds of millions of pages via its crawlers. But at the time of the report in mid-October 2026, the question of whether this indexing generates visits to the sites remained open.
Google, on its part, has deployed several updates to make sources more clickable in its AI responses. Links are becoming more visible in AI Overviews, suggesting that the engine has measured the imbalance. Following the news on Declic Web helps keep an eye on these developments over the weeks.
In practical terms, we can no longer be satisfied with checking if content is “cited” by AI. The relevant metric becomes the actual click-through rate from the generated response, and most standard analytics tools do not yet isolate it. It is necessary to cross-reference data from the Search Console with server logs to identify visits that can be truly attributed to conversational interfaces.
Measuring the Value of an AI Citation When Clicks Are Missing
A site that appears in a response from Perplexity, ChatGPT Search, or an AI Overview finds itself in an unprecedented position: its brand is exposed without the traffic following. We talk about visibility without direct conversion, somewhat like a billboard where we cannot count the passersby who push the door.
Three Ways to Evaluate This Presence
- Monitor brand queries in the Search Console. If the site’s name appears more often as a direct query after an AI citation, it is an indirect signal of acquired notoriety.
- Compare the pages cited by AI with their evolution in classic organic traffic. A cited page that loses standard Google traffic but gains in brand queries is undergoing a transfer of value, not a disappearance.
- Analyze server logs to identify the user agents of AI crawlers (GPTBot, ClaudeBot, PerplexityBot) and measure the crawl frequency per page. An intensive crawl without associated traffic signals content being exploited without return.
The limitation of this approach lies in the granularity of the available data. Conversational assistants do not all transmit an exploitable referrer, and the share of “dark traffic” (visits without an identified source) is mechanically increasing.
European Regulation: the DSA Catches Up with AI Engines
The European Union has classified ChatGPT as a very large search engine under the Digital Services Act (DSA). This classification entails additional obligations announced for January 2027, particularly regarding algorithmic transparency and result moderation.
The distinction is technical but has significant consequences. When ChatGPT provides answers that include links and excerpts from third-party sites, it no longer operates as a simple chatbot. It acts as a search intermediary, just like Google or Bing.
For site publishers, this opens a concrete perspective: platforms classified as “very large engines” will have to account for how they select, display, and attribute content. We are not yet at detailed audits, but the legal framework now exists.
What the DSA Changes for a Publishing Site
A publisher who finds that their content feeds the responses of an AI engine classified under the DSA has a regulatory lever. They can request information on the selection criteria and, ultimately, contest a lack of attribution. This mechanism did not exist before this classification.
Web Accessibility: The New WCAG Criteria Changing Front-End Development
Web trends are not solely about AI. Accessibility requirements are evolving within the European framework, and the criteria added by the WCAG 2.2 update directly impact the daily work of front-end developers.
Among the notable additions: criteria concerning keyboard focus visibility, mandatory alternatives to drag-and-drop, and requirements for the minimum size of touch targets. A button that is too small on mobile, a menu accessible only by drag-and-drop, an invisible focus indicator: these common flaws are becoming non-compliances.
The WCAG 3.0 standard remains a work in progress and does not replace the 2.x versions. Therefore, we are working with WCAG 2.2 as the operational reference, anticipating that the next version will likely tighten the criteria on dynamic interfaces and components generated by JavaScript.
AI Bots and Content Protection: A Sharp Increase in Traffic
The traffic from malicious bots has increased dramatically in recent months. The term “open web” does not mean that all published content is freely exploitable by any crawler. Publishers are facing a multiplication of bots that scrape content without respecting robots.txt files.
On the ground, the response involves behavioral detection tools (DataDome, Cloudflare Bot Management) capable of distinguishing legitimate crawls from aggressive scraping. The robots.txt file remains a declarative signal, not a technical barrier. A site that publishes valuable content should monitor its logs and block unwanted agents at the server level.
The digital news at the end of 2026 converges toward the same conclusion: the value of web content is increasingly measured not just by incoming traffic. AI citation, regulatory compliance, and the ability to protect its pages against automated scraping are becoming indicators as concrete as bounce rates or page views.
