From Blue Links to Agentic Search
Google is turning its classic search box into an AI front-end, where “Google Search Is AI Search” is now a product strategy, not a slogan. At I/O, Search chief Liz Reid described a system that can act, not just rank pages: Gemini agents will plan, retrieve information and perform multi-step tasks directly in the results. The upgraded box, powered by Gemini 3.5 Flash, expands according to query type and merges previously separate AI experiences like AI Mode and AI Overviews into a single, conversational interface. For publishers, this means the primary experience increasingly happens inside Google’s environment, not on their sites. With AI Overviews already influencing behaviour at scale and Gemini usage measured in the hundreds of millions monthly, the shift signals a structural change in how people ask questions, see answers and discover content across the web.
Multi‑Step AI Agents and the Vanishing Click
The most disruptive change for publisher visibility is the arrival of Google AI agents in Search. These agents can execute multi-step retrieval and planning: researching options, comparing information, and even tying into tools like Gmail or Calendar to act on a user’s behalf. Instead of serving a page of ranked links, the interface now leans toward synthesized, assistant-style answers that keep users inside Google’s conversational search features. TechRepublic notes that this Gemini search integration already encourages prompt-like queries and seamless switching into AI Mode. SEO experts warn this will “undoubtedly lead to fewer clicks,” building on patterns seen with AI Overviews where a majority of users either refine the query or stop browsing rather than click a traditional result. As agents become better at finishing tasks end-to-end, publishers risk being reduced to background data sources rather than destinations.
How Gemini 3.5 Flash Changes Search Results Ranking
Gemini 3.5 Flash sits at the core of these AI search results ranking changes. It allows the search box to interpret natural-language prompts, understand context over follow-up questions and handle multimodal inputs like images, videos, files and URLs. Instead of ranking pages purely on keyword relevance and links, Google increasingly ranks “experiences”: coherent, conversational responses stitched together from multiple sources. AI Overviews and the integrated AI prompt box encourage users to refine their intent through dialogue, giving Gemini more behavioural data to shape results over time. That feedback loop may prioritise content that is easier for models to parse, summarize and reuse. For publishers, it means traditional SERP positions matter less than how well their content can be ingested by Gemini, surfaced within summaries and reused in agentic workflows that users may never see as a list of discrete websites.
The New Reality for Publisher Visibility and Traffic
Evidence suggests the visibility problem is already here. A Pew study cited by TechRepublic found that when an AI Overview appears, only 8% of users click a traditional link, versus 15% when no overview shows. Many users instead reformulate queries, navigate directly, or simply stop. SEO consultants report that some sites now see the majority of their informational queries served primarily for AI overview purposes. Industry voices expect more people to engage with the web through AI interfaces, spending less time on individual sites in their browsers. While some draw optimistic parallels to past shifts like mobile SERPs, others question whether total clicks can grow when AI is systematically interposed between user and publisher. The net effect is a squeeze on publisher visibility SEO: less exposure per query, weaker brand recall, and greater dependence on how Google’s AI chooses to represent their work.
Adapting SEO Strategy for an AI‑First Search Landscape
To survive in an environment where Google AI agents search, summarize and act, publishers must optimize for being the best possible input to Gemini, not just the highest blue link. That starts with strong E‑E‑A‑T signals: original reporting, unique data, clear author credentials and transparent sourcing that help systems identify trustworthy expertise. Structured data becomes essential, giving Gemini machine-readable context about entities, events, products and relationships so it can confidently quote or surface content in synthesized replies. Direct answer optimization—concise explanations, definitions and step-by-step instructions—improves the odds of being cited within conversational search features, even if the user never visits the page. Publishers should also track how their content appears inside AI Overviews and agentic flows, experiment with more niche or depth content less exposed to summary cannibalization, and diversify beyond search to reduce dependency on a single AI-driven channel.