Navigating Google’s AI Overviews: New Link Dynamics, Tracking Challenges, and SEO Implications
As Google continues to weave artificial intelligence into the fabric of its core search engine, the mechanics of how users discover, interact with, and navigate web content are undergoing a seismic shift. Recent industry data, user observations, and updates to Google Search Console have thrust the spotlight onto a complex web of link behaviors inside AI Overviews and AI Mode.
For digital marketers, publishers, and SEO professionals, understanding these changes is no longer optional—it is vital for accurate performance measurement. Not every link inside an AI-generated response functions the same way, and the gap between what users see on their screens and what analytics report in Search Console creates a challenging environment for traffic attribution.
The Main Facts: The Evolving Landscape of AI Search Links
Google’s deployment of generative AI in search has rapidly evolved from a novel experimental feature into a foundational interface layer. However, recent findings highlight a growing divergence between traditional outbound links and interactive elements designed to keep users within the Google ecosystem.
Recent observations from the SEO community reveal two distinct phenomena: a sharp upward trajectory in external links embedded directly into AI Overview text, and the proliferation of anchor links that bypass external websites entirely, funneling users directly into conversational "AI Mode."
According to Google’s documentation, the definition of a "link" inside an AI overview varies wildly. Links can appear as text citations next to an answer, article suggestions at the conclusion of a response, interactive previews of online discussions with community labels, or—increasingly—conversational follow-up prompts designed to deepen engagement.
Yet, beneath these visual updates lies a complex tracking reality. In Google Search Console, clicking an external link within an AI Overview is registered as a click. Conversely, interactive links that trigger a new internal Google query (such as a query refinement) do not register as clicks or impressions for external publishers, fundamentally altering how traffic flows are measured.
Chronology of Google’s AI Link Updates (2025–2026)
To understand how we arrived at the current state of AI search metrics, it is helpful to trace Google’s rollouts and executive statements over the past year:
- August 25, 2025: Robby Stein, VP of Product at Google Search, stated that Google’s long-term "north star" was to display significantly more inline links within AI-driven search results, introducing desktop link carousels.
- January 27, 2026: Google rolled out global mobile support allowing users to click follow-up questions directly inside AI Overviews, seamlessly transitioning users into a conversational back-and-forth within AI Mode.
- February 17, 2026: Stein announced forthcoming desktop UI updates, including pop-up link groups on hover and more descriptive link icons for both mobile and desktop platforms.
- May 6, 2026: Hema Budaraju, VP of Product Management for Search, detailed five major updates to AI Overviews and AI Mode. These included end-of-response article suggestions, labeled subscription links, public discussion previews, and a higher frequency of inline links.
- June 3, 2026: Google confirmed an increase in inline links within AI responses, noting ongoing experiments with novel UI designs.
- June – August 2026: Google tested and subsequently rolled out a dedicated Generative AI performance report in Search Console, expanding availability worldwide by August 31, 2026.
- September 2026: Industry analysts like Peec AI and SEO practitioners like Gagan Ghotra published data and video evidence highlighting sudden surges in external link placements alongside new AI Mode anchor link tests.
Supporting Data: What Analytics and Trackers Reveal
Recent data shared by analytics platforms and industry experts provides a window into how these interface updates manifest in the wild.
The Rise of External Links in AI Overviews
On September 21, 2026, Malte Landwehr of AI search analytics firm Peec AI published a chart illustrating a dramatic surge in external links embedded directly within AI Overview text. Tracking daily averages from August 25 to September 20, Peec AI observed that the share of AI Overviews containing an in-text external link hovered near zero for weeks.
However, on September 11, the metric experienced a sharp vertical spike, steadily climbing to reach 26.2% by September 20. While the underlying dataset lacked granular breakdowns regarding specific markets, devices, or exact query volumes, the data strongly suggests a deliberate algorithmic or UI adjustment by Google to increase the prevalence of web citations inside generative answers.
The AI Mode Anchor Link Phenomenon
Earlier that same day, SEO practitioner Gagan Ghotra shared an 18-second video capture demonstrating a newly observed behavior. When clicking on bold, underlined text inside an AI Overview—traditionally expected to point to an external editorial source—the user was instead redirected into an active AI Mode conversational prompt.
This test highlights an ongoing balancing act by Google: providing pathways to the broader web while aggressively training users to adopt conversational, AI-driven search loops.
Official Responses and Search Console Mechanics
As publishers grapple with these shifting dynamics, Google’s Search Console help documentation provides specific guidelines on how traffic and visibility are calculated within generative features, though notable ambiguity remains.
How Clicks and Impressions are Counted
Google Search Console treats interactions based strictly on destination:
- External Page Clicks: Clicking a link that successfully navigates a user to an external website inside an AI Overview counts as a standard click.
- Query Refinement Links: If a user clicks a link that executes a brand-new Google query (such as a topical drill-down), Search Console does not count impressions or clicks for that specific link. Instead, the subsequent search results page is treated as an entirely fresh query.
- The AI Mode Gray Area: While official documentation clearly outlines how follow-up questions in AI Mode function as new queries, it remains officially unspecified whether the specific inline anchor links demonstrated by Ghotra are classified as query refinements. If they are, they yield zero attribution data for external site owners.
The Generative AI Performance Report
The rollout of the dedicated Generative AI performance report on August 31, 2026, provided site owners with isolated impression data for AI Overviews and AI Mode, filterable by page, country, date, and device.
Crucially, this report measures impressions only—it does not list clicks or click-through rates (CTR). Because Web search performance metrics pool clicks from both traditional results and AI features together, publishers cannot mathematically derive a definitive CTR for AI Overviews alone. Furthermore, chart totals aggregate impressions by property; if multiple pages from the same domain appear in a single AI Overview, they register as a single property impression unless filtered by URL.
Implications for Publishers and SEO Strategies
The intersection of rising AI Overview external link frequencies and the expansion of AI Mode navigation presents profound implications for digital publishing and search engine optimization.
1. The Attribution Gap
The inability to isolate clicks originating exclusively from AI Overviews makes traffic forecasting exceptionally difficult. Publishers may observe climbing impressions in the new Search Console AI report, yet remain entirely blind to whether those impressions are translating into actual referral traffic or simply feeding passive user glances.
2. Behavioral Conditioning
Industry analysts speculate that Google’s interface updates are actively training users to interact differently with search results. By interspersing external reference links with conversational AI Mode triggers, Google creates a mixed ecosystem where the boundary between a traditional web destination and an internal AI chat prompt becomes increasingly blurred.
3. Adapting Content Strategies
For SEO teams, optimizing for AI Overviews requires moving beyond traditional keyword targeting. Because Google is experimenting heavily with inline links, subscription callouts, and discussion previews, content must be structured to provide authoritative, easily summarizable insights that AI models can comfortably cite as inline sources. However, publishers must also prepare for a scenario where user engagement increasingly happens around the content rather than on it.
Looking Ahead
The rapid iteration of Google’s search interface signals that generative AI features will continue to evolve rapidly. While Search Console’s dedicated AI reports have brought much-needed visibility to impression metrics, the ongoing lack of granular click attribution leaves critical questions unanswered.
Future clarity will likely depend on expanded reporting features from Google, coupled with longitudinal studies from third-party analytics trackers capable of mapping the complex journey from AI Overview impression to eventual site visit. Until then, publishers must monitor their aggregate performance metrics closely, remaining adaptable as Google refines the delicate balance between answering queries directly and driving traffic to the wider web.
