The arrival of AI curators is rewriting the rules of online discoverability. What used to be a predictable pipeline , write, optimize for organic rank, get clicks , is now being interrupted by automated summaries and recommendations that answer users without sending them to your site.
Major platforms and studies show this is more than a trend: Google calls AI Overviews one of its most successful Search launches in a decade, and independent analyses show AI summaries appearing on an increasing share of informational queries. Bloggers who ignore AI curators risk losing referral traffic, attribution, and long-term visibility.
The rise of AI Overviews and the curatorial layer
Google’s May 20, 2025 update describes AI Overviews and a new “AI Mode” powered by Gemini 2.5 as driving meaningful increases in Search usage, and the company reports AI Overviews boost usage on eligible queries by over 10% (Google, May 2025). This product-level shift signals that AI curators are now an integral discovery layer on top of classic search results.
Independent measurements confirm rapid adoption. Semrush found AI Overviews appearing on about 6.5% of queries in January 2025 and rising to over 13% by March 2025, mostly for informational searches. That concentration on informational intent matters because those are the queries where blogs and long-form posts historically gained the most referral traffic.
At the same time, AI curators do not simply mirror the old top‑10. Semrush’s large analyses show AI Overviews often cite pages outside the canonical rankings, and mobile AIOs sometimes include three or fewer top‑10 results. For creators, that means the new curator layer surfaces different signals than classic SEO.
Zero‑click patterns, impressions up, clicks down
Data from BrightEdge reported in May 2025 reveals a striking pattern: Google search impressions rose roughly 49% year‑over‑year, while overall click‑through‑rates fell about 30%. BrightEdge attributes much of that CTR decline to AI Overviews, where users consume summaries instead of clicking through to sources.
Publishers have reported real losses. Coverage aggregated by outlets such as The Washington Post and publisher surveys show some sites experienced steep traffic declines after AI summaries were introduced; Raptive estimated potential creator revenue losses in the order of billions. Those effects underscore a harsh reality: higher visibility in search impressions no longer guarantees audience visits.
For many blogs the metric that matters is shifting. Inclusion in AI Overviews , being cited or recommended by an AI curator , is becoming a primary reach KPI, because it determines whether readers are directed to the original content or served a synthesized summary inline.
Volatility of AI citations and why continuous monitoring matters
AI curators are volatile. Semrush’s URL volatility study found that around 96% of AI Overviews showed domain changes over a month and that approximately 91% of URLs were removed from AIOs at some point. Only about 43% of removed URLs returned within the same month, illustrating unstable citation behavior compared with classic organic rankings.
That volatility means discoverability is no longer a set‑and‑forget problem. Where classic SEO rewarded sustained ranking efforts, AI curators can quickly reshuffle which pages are cited. Continuous monitoring, rapid content updates, and diversified distribution become essential tactics for blogs that want steady reach.
Practically, volatility also raises questions about attribution, provenance, and trust. If AI curators rephrase or aggregate content without stable citation, creators lose both traffic and credit , a threat that underscores the need for better metadata and clearer provenance signals.
Practical optimizations: GEO, metadata, and provenance
Industry recommendations and platform guidance converge on a new set of tactics sometimes called Generative Engine Optimization (GEO). Google’s Search Central updated guidance (May 21, 2025) urges authors to focus on unique, valuable content for people, ensure technical accessibility, use structured data properly, control snippet behavior when needed, and improve page experience , all to increase the chance of success in AI experiences.
Other GEO tactics include explicit citation and attribution markup, clear author/publisher provenance, strong E‑E‑A‑T signals, and structured metadata that helps AI curators identify original sources. Academic work on data curation and provenance recommends watermarking and richer provenance metadata so human‑created content is easier for automated curators to trace and credit.
Because AI Overviews often cite pages beyond the top‑10, optimizing for these curator signals , rather than only chasing rank , can help niche and long‑tail posts surface. In short: SEO is no longer just about ranking , it’s about being recommended and cited.
Business and distribution responses publishers are testing
Publishers and creators are experimenting with multiple business responses to reduced referral traffic. Some are moving content behind paywalls or APIs and negotiating licensing deals with AI platforms so their content can be included on compensated or credited terms. Others are productizing newsletters, memberships, and direct channels that reduce dependency on search referrals.
At the same time, a growing ecosystem of reader apps and hybrid curation services is emerging. Indie players and reader apps are adding AI summaries and recommendations (for example, Refind’s hybrid human+AI model), offering alternative discovery paths that can bypass both traditional search and large‑scale AI Overviews.
Those platform shifts create both threats and opportunities: blogs might lose some organic search clicks but can gain audiences via curated feeds, partnerships, or licensing , if they prepare content so it’s visible and creditable to those curators.
Preparing for agentic AIs and the next discovery frontier
Academic research now frames a longer‑term shift: agentic AIs that autonomously browse and act on the web will require a new discipline (sometimes called Agentic AI Optimisation, or AAIO). An April 2025 arXiv paper argues that autonomous agents change discoverability dynamics and that sites must proactively design content and governance to remain findable and correctly attributed.
Gartner’s widely cited prediction that traditional search‑engine volume could fall about 25% by 2026 because users shift to AI chatbots and agents amplifies the urgency. For blogs that depend heavily on search referrals, that projection is a strategic risk signal to diversify reach and adopt agent‑aware practices now.
In practice, the future favors creators who build stronger provenance metadata, improve technical accessibility, maintain up‑to‑date structured data, and pursue partnerships with curators and platforms. These actions make content easier for automated curators to find and credit , a competitive edge as AI curators evolve.
AI curators are reshaping how readers discover, consume, and credit online writing. The combination of product change (Google’s AI Overviews), independent measurements (Semrush, BrightEdge), academic warnings (AAIO and data curation papers), and publisher losses (reported industry estimates) makes clear that discoverability is becoming curator‑centric rather than rank‑centric.
For bloggers the takeaway is pragmatic: adopt GEO tactics, improve provenance and structured metadata, diversify distribution, and monitor AI citation behavior continuously. If you want the source PDFs or direct links to the Semrush, BrightEdge, Google, and arXiv studies cited here, I can pull and list them for deeper reading.
