The boom of AI-powered search engines is fundamentally reshaping how users access information on the Internet. According to a recent VentureBeat report citing Pew Research Center, users click on traditional search results only about 8% of the time when Google displays AI Overviews, nearly halving the typical 15% click-through rate. More notably, links cited directly within AI answers generate a mere 1% click-through rate.
Context & Drivers
This decline has dealt a heavy blow to the ad-based business models of publishers. Data from Chartbeat shows that Google Search traffic across the publishing ecosystem fell by up to 34% between late 2024 and late 2025. Small publishers saw an even more severe drop of up to 60%.
This trend reflects a reality where AI keeps users within its chat interface rather than directing them to original websites. However, behind this decline lies a paradox: AI models are actually crawling and consuming web data at a breakneck pace to improve the quality of their responses.
Technical Analysis & Technology
Technical data from Similarweb shows that the proportion of ChatGPT answers containing direct web citations increased more than fivefold, reaching 6.8% by May 2026. In niche sectors like travel, this rate reached as high as 22.6%. Instead of directing users to deep pages containing in-depth articles, AI tends to send them straight to the homepage.
When ChatGPT updated its feature to display prominent brand links, referral traffic from the chatbot skyrocketed by 157% in just one week. Crucially, the share of traffic directed to homepages surged from 25% to nearly 60%. This shift occurs because the AI model handles all information synthesis and comparison, leaving users to visit the homepage only when they are ready to make a purchase or sign up.
Expert Perspectives & Insights
According to Lily Ray, Vice President of SEO and AI Search at Amsive, a website's visibility on traditional search engines directly dictates its chances of being cited by AI models. If organic visibility drops, AI systems will fail to discover and synthesize your data. An analysis from Ahrefs points to a major challenge for marketers: up to 67% of the most-cited sources by ChatGPT lie outside the control of marketing campaigns, with Wikipedia alone accounting for nearly 30%.
Impact & The Future
This paradigm shift is forcing businesses in Vietnam and globally to urgently restructure their websites to adapt to the new era of search. Deep content pages need to be clearly structured in natural language to facilitate easy crawling by AI, while homepages must be redesigned to cater to high-intent users who already have purchase context.
Finally, internal search engines—long neglected by web developers—need a major upgrade. An analysis by Previsible reveals that up to 28.8% of referral visits from ChatGPT land directly on these internal search pages. The era of the traditional click is drawing to a close, giving way to a race to optimize for both human users and machine learning algorithms.