Watching clicks vanish while AI answer boxes steal the stage? This roundup decodes the quiet upgrades reshaping search that include image-powered answers, premium AI tiers, and bots that pay to crawl. Expect rapid headlines, data you can benchmark, and actions to protect visibility. Ready to see the shifts? Scroll on. Let’s get started.
1. Rapid-fire updates reshaping the generative search landscape
A burst of product launches, policy fights, and traffic shocks rattled GEO this week. Scan the headline moves below to spot what hit your roadmap fastest.
- Google — AI Mode now shows up inside Circle to Search, letting Android users tap an image and ask follow-ups without leaving the screen. (Google)
- Perplexity — New “Perplexity Max” tier rolled out on 2 July with unlimited Labs and early access to Comet. (Perplexity )
- Anthropic — Claude API adds search result content blocks so responses carry first-party citations by default. (Anthropic)
- Independent publishers — Filed a formal antitrust complaint in Brussels, arguing Google’s AI Overviews hijack traffic. (Reuters )
- EU policy — Forty-five tech firms asked regulators to postpone the AI Act’s GPAI obligations slated for August. (Reuters )
- LiveRAG — SIGIR’s challenge recap shows 70 teams across 27 countries racing to answer 500 questions in two hours. (arxiv)
- Traffic shift — Bot crawling is up 125% in six months, squeezing publisher referrals. (Reuters)
Also read: Infrastructure Shake-Up in Generative Engine Optimization
2. Inside stories driving the biggest shifts in generative discovery
From mobile image answers to premium AI tiers, these deeper narratives explain why the search stack feels different already. Explore the key stories below.
Circle to Search gets AI Mode
Google just put its full-fat AI answer engine inside the gesture-based Circle to Search flow. Mobile shoppers can highlight a product photo, press “dive deeper,” and receive an overview plus linked sources without opening Chrome. That widens the surface where snippets replace clicks; brands that depend on image search now need structured data ready for follow-up questions. (Google)
Perplexity Max Targets Power Users
The new $35/mo “Max” plan unlocks unlimited Labs runs, priority support, and frontier models like o3-pro and Claude 4. Early adopters gain first crack at Comet, Perplexity’s side-panel browser. Expect answer volume from Max users to spike, shifting citation demand toward sites that load quickly and carry clear facts. (Perplexity)
Publishers vs. Google Heat Up in the EU
A coalition led by the Independent Publishers Alliance asked the Commission to impose interim measures against AI Overviews, claiming abuse of dominance and lost ad income. The case puts source attribution and opt-out controls on the legal agenda just as AI Mode expands. GEO teams should prepare discovery files that document traffic swings — regulators will cite them. (Reuters)
LiveRAG Challenge Highlights The Retrieval Edge
Seventy teams tackled a 15M web document corpus using identical Falcon3-10B weights; top scores clustered around retrieval tweaks, not prompt art. That’s a reminder that ranking signals still decide which URL gets quoted when models tie on generation quality. RAG evaluation prizes ($5k for first place) show there’s now money — and benchmarks — for micro-improvements in recall. (arxiv)
3. New numbers showing how AI answers are rewiring traffic flows
Fresh data quantifies the swing from human clicks to bot activity, and what it costs. Check out the metrics below to gauge where you stand.
- Zero-click news queries on Google rose from 56% to 69% year-over-year after the AI Overviews launch. (Pressgazette UK)
- Mail Online keywords now trigger AI Overviews 68.8% of the time; 69 % of those sessions end without a click. (Pressgazette UK)
- LiveRAG’s finals capped at 70 teams / 27 nations / 500 Qs / 120 min, underscoring global RAG talent density. (arxiv )
- Cloudflare handled “hundreds” of publisher interviews before shipping Pay Per Crawl, signalling market demand for paid access*.* (Cloudflare)
- Bot traffic to content sites jumped 125% in six months, vs. single-digit growth for human visits. (Reuters)
Also read: Multimodal Content Dominates AI Search
4. Lesson from Mail Online on surviving an era ruled by answer boxes
Searchers stop clicking—thanks to AI. One giant’s stumble reveals how zero-click results drain even top brands. Check the details below before the same pattern hits you.
- What changed? Similarweb data shows Mail Online’s zero-click searches for its top 100 keywords climbed from 48 % (May '24) to 68.8 % (May '25) as AI Overviews spread. (Pressgazette UK)
- Evidence: Average organic visits fell from 2.3B to under 1.7B in the same period. (Pressgazette UK)
- Takeaway: High-profile domains are no longer safe; optimise snippets for extraction and push schema that nudges models to cite — or risk living in a zero-click world. (Pressgazette UK)
5. GEO Tactic of the Week — Claude Search Blocks
Claude’s new search result content blocks let developers pass structured web hits (title, URL, snippet) directly into the prompt; Claude then cites them automatically. No more juggling custom footnotes. Ideal for marketers building internal GEO dashboards or FAQ bots. Roll out by adding the beta header search-results-2025-06-09. (Anthropic)
Also read: 5 Game-Changing GEO Trends That Will Define Search in the Next 90 Days
6. What Steps Can You Take to Ensure You Have an Advantage in GEO?
- Label your images and alt-text — Circle to Search now reads pictures; feed it product specs and prices in structured data.
- Audit retrieval quality — Use LiveRAG’s public dataset to benchmark your RAG stack; focus on recall, not model swaps.
- Article schema — Treat every headline, deck, and image as structured input—surface crisp fact blocks, authorship, and dates in Article schema. Add speakable tags so answer engines can lift your content accurately and credit you as the original, authoritative source.