AI search intelligence — daily analysis of how AI engines are changing discovery, what it means, and who is adapting.
If OAI-SearchBot traffic spiked this month, blocking it is the wrong move: that costs ChatGPT search eligibility. Five weeks of logs across five domains show why — ChatGPT-User fetches fell 49% while OAI-SearchBot crawls rose 838%, and 87% of that crawl rise came from one domain, a site event, not an engine change.
The Machine Relations Index now publishes per-engine run denominators. Perplexity cited a source in 99.97% of returned runs. Gemini cited one in 48.4%.
Machine Relations' live Fidelity cohort has now graded 8,258 ChatGPT citations at 53.4% supported, confirming the September frozen-sample rate of 54.8%. Perplexity, Gemini and Claude still have no live-cohort rate.
Across 14,447 readable citations in four AI engines, the cited source's name reached the answer text 8.6% of the time. The rate splits by engine.
The Machine Relations Index's raw exclusivity share for all six AI answer engines, then the same number conditioned on citation volume. Among the three engines the Index collected without a gap, ChatGPT and Perplexity keep an elevated exclusive-source share past 5 citations and Claude collapses to 1.1%. The three gapped surfaces cannot be ranked on this measure.
Our 67-domain reach gap between Google AI Mode and AI Overviews is withdrawn: the instrument had collection gaps. The surfaces share 363+ of the top 500.
Every AI visibility tool reports domains cited. Almost none tells you whether learn.microsoft.com counts as microsoft.com. The Machine Relations Index publishes its rule and tested it on every published row: it holds on 23,262 of 23,280, and the 18 edge cases show exactly where any tool's domain count drifts.
A dated record of what Claude cites. Entry one: Claude reaches 463 of the 500 most-cited domains, second only to Perplexity among the engines measured.
A dated record of what Perplexity cites. Entry one: Perplexity reaches 490 of the 500 most-cited domains, the widest head overlap of any engine measured.
An 89-day first-party panel of 35 queries across five AI surfaces: when an engine keeps a domain, it cites the exact same URL 72% to 91% of the time.
A dated record of what Gemini cites. Entry one: Gemini reaches at least 452 of the 500 most-cited domains, above ChatGPT on every composition measure.
The first published HR software buyer-choice leaderboard looks stable at the category layer and fragmented at the engine layer. Four of its top ten sources are absent from at least one of the three engines the Index collected without a gap in this window, while the HR news list is universal across all three.
Reddit ranks #1 of 22,320 domains in the Machine Relations Index, but its engine breadth is 4 of 6: ChatGPT and Claude show zero citations in the same window.
Measured against 15,883 monitored answer runs, wikipedia.org ranks #27 of 22,213 cited domains at a 1.22% citation rate. Its rate ranges from 11.43% to 0.74% depending on the question shape, and it is observed in 17 of 25 released categories.
A calibrated instrument graded 15,525 citation events across four AI answer engines. ChatGPT citations support their claim 54.8% of the time, Perplexity 50.1%; Gemini and Claude publish no rate. One failure mode is publisher-side: the cited page, fetched back minutes later, could not be read at all.
Reddit is the most-cited domain in AI answers and is absent from ChatGPT and Claude. The per-engine map of the top 25 sources, what each engine's robots and licensing layer says, and what changed since July.
A dated record of what ChatGPT cites and what changes, read from the Machine Relations Index. Entry one: ChatGPT has the weakest overlap with the index head of any of the six engines, reaching 356 of the top 500 where Perplexity reaches 489.
Use missing-answer sensitivity bounds to test whether an AI citation leader survives collection gaps, without turning failed requests into measured zeros.
AI citation dashboards should assign daily rates by observation time, record ingestion time separately, and publish versioned restatements for late arrivals.
AI citation-share dashboards should separate successful answers with no citations from failed, timed-out or unscheduled observations before reporting a rate.
AI citation-rate reports should separate scheduled observations, retry attempts, successful answers and repeated URLs before publishing any denominator.
AI citation-rate reports should version prompt cohorts, separate retained prompts from additions and removals, and label full-basket rates noncomparable.
AI visibility teams should treat RAG citation share as the control plane: model-layer brand mentions stay volatile, cited source selection is verifiable.
AI citation reporting should preserve the URL an answer cited and record its later redirect destination separately, with timestamps and status codes.
Google AI Overviews image generation makes visual source architecture a search visibility problem.
OpenAI separates ChatGPT Search, training crawls, and user-triggered fetches into different access rules.
ChatGPT connectors move AI search from public web retrieval into private enterprise source routing.
Google's new Flash model shows why agentic search favors fast, citable source architecture.
AI citation rate is replacing rank as the practical unit for measuring AI search visibility.
Meta crawler docs make AI search visibility a crawl, source, and citation architecture problem.
Perplexity's Time ad block shows AI search trust is becoming source governance.
Google AI Search controls make publisher consent a source infrastructure problem.
Cloudflare Radar Researcher shows why AI search now needs live, queryable source infrastructure.
Cloudflare is making retrieval an agent primitive, not a bolt-on search feature.
The acronym fight matters less than the stack: crawl, extract, cite, and measure.
Google Ask Maps makes local discovery personalized, agentic, and harder to measure with rankings alone.
AWS Bedrock Web Search makes grounded AI answers a source-architecture problem.
Google AI Search now rewards source control, not just rankings. Brands need machine-readable authority.
Onton's benchmark signals that AI search rewards structured product evidence over keyword matching.
Reddit's AI Overview warning shows AI search still depends on source traffic.
OpenAI's 80% cut on GPT-5.6 Luna signals a model price war that keeps resetting which brands AI search cites.
AI Overviews now show on 43% of Google searches and most sessions end without a click. Visibility now means being cited inside the answer.
Google's always-on Gemini Spark agent is rolling out worldwide. When an agent does the searching, ranking on the SERP stops mattering.
Claude shared chats got indexed by Google. The real signal: AI share links are now an indexable content class.
Gemini 3.6 Flash and Claude Opus 5 both became default models in one week. When the model changes, so does what it cites. Here's why.
AI referrals are outgrowing GA4 defaults. Attribution now needs crawler, referrer, and citation-level proof.
PR for AI search is shifting from media outreach to answer-engine source architecture.
PRWeek's AI-search report shows PR content is becoming source architecture for answer engines.
OpenAI Presence turns voice and chat agents into managed retrieval surfaces for enterprise brands.
Salesforce's agentic commerce search shows product discovery shifting from keywords to intent graphs.
Google's AI Search click claim shifts the open-web fight from traffic volume to source architecture.
New data across 137,000 domains shows 97% of llms.txt files never get read. Here's what actually moves AI citations.
Perplexity WANDR shows AI research agents still fail at evidence depth, not just page discovery.
Germany's ZAK ruling reframes AI search answers as accountable media surfaces.
Google AI Mode connected apps make search an execution surface, not just an answer surface.
GA4's AI Assistant channel is useful, but it is only the visible floor of AI search traffic.
Google Images is becoming a visual discovery surface instead of a static image archive.
A Bocconi clickstream study finds ChatGPT refers out in 5.2% of sessions vs Google's 31.1% — and cuts traditional search 17% in twenty weeks.
A 320-query study shows Google AI Mode and Perplexity cite by SEO rank, while ChatGPT and Claude ignore it. One playbook no longer works.
Anthropic's Jacobian Lens shows why output-only AI visibility dashboards measure too late.
Muck Rack's new PR study shows AI search visibility has an ownership gap, not just a tactics gap.
AI answer engines are adding native checkout. What agentic commerce changes for brand visibility—and who gets bought.
OpenAI's GPT-Live makes ChatGPT Voice a search surface with web search, memory, and visual answer cards.
Cloudflare and OpenAI are testing whether AI search can crawl by freshness signals instead of blind recrawling.
No AI engine publishes a weight for earned media. What each documents instead, with dates: Google, Bing, Perplexity, ChatGPT and Claude as of October 3, 2026.
OpenAI documents automatic chatgpt.com UTM tagging on ChatGPT referrals, which makes AI clicks visible in analytics while citation exposure stays unmeasured.
A field experiment shows AI Overviews cut publisher clicks. The new signal is citation, not rank.
Pinecone Nexus signals enterprise search moving from document retrieval to agent-ready knowledge layers.
Web Bot Auth moves AI crawler access from user-agent claims to signed identity.
Cloudflare's AI search push turns freshness, permissions, and usage data into ranking infrastructure.
Claude Fable 5 is back, but the real signal is model access becoming AI discovery infrastructure.
Claude in Microsoft Foundry turns enterprise AI retrieval into source-architecture work.
OpenAI's finance rollout turns ChatGPT into a high-intent vertical answer surface.
Google's reported Gemini limits on Meta show AI search is becoming a rationed control plane.
Google's June 2026 spam update turns AI content quality into a source architecture problem.
Seltz's seed round shows agent search is becoming a machine-readable infrastructure layer.
GEO manipulation is moving from SEO tactic to AI search security risk as answer engines absorb web evidence.
Claude Tag makes Slack a shared AI answer surface, forcing brands to rethink internal source architecture.
Google's Interactions API makes Gemini agent workflows a source-architecture problem.
Google's new generative AI reports move AI visibility from guesswork toward source-level measurement.
Google and Microsoft-backed ARD turns AI agent discovery into a protocol and source architecture problem.
Meta's AI Mode turns Facebook search into a synthesized answer engine pulling from public posts. The blue-link era of social discovery just ended.
Bing Webmaster Tools now reports Citation Share, Intents, Topics, and Compare. The AI-visibility yardstick just moved from binary to market share.
A Munich court held Google directly liable for false AI Overviews claims — the first ruling to treat synthesis as authorship, not search.
AI Overviews broke the link between rank and clicks. A real AI visibility audit measures citation across engines, not position. Here is the framework.
Perplexity now routes Deep Research across 20+ models. Your citation is no longer one model's call — it's an orchestration outcome.
AI Overviews now run ads inside the answer box. The data shows clicks aren't collapsing — Google is repricing them. Here's what actually changed.
Google's 'new era for AI Search' is live globally. Organic links are now a side panel. Citation, not ranking, is the discovery surface that remains.
Microsoft Web IQ is a search engine for AI agents, not humans. Passage-level retrieval just made the webpage the wrong unit of discovery.
Anthropic's hidden Claude guardrail shows why AI search needs source and fallback disclosure.
SparkToro's new zero-click data shows why AI Mode makes citation presence more important than traffic.
Google's NotebookLM update shows source discovery moving from search results into AI research chat.
Apple's Siri AI turns assistant answers into a new retrieval layer for brands.
OpenAI's Lockdown Mode shows agentic search now has a security-versus-freshness tradeoff.
OpenAI's memory upgrade makes AI search more persistent, personalized, and harder to measure.
Yahoo Scout is moving AI answers into Sports and Finance, turning vertical archives into retrieval products.
Google's AI Search controls turn publisher inclusion into a retrieval-governance decision.
Perplexity Search as Code shifts AI search from ranked results to executable retrieval workflows.
Google's AI answer layer is aligning faster with Search enforcement. That changes AI visibility risk.
Google's Universal Cart turns AI search into a commerce control layer, not just a shopping feature.
AI search now rewards extractable evidence, not generic content volume.
Agentic search is turning web infrastructure into a retrieval surface. Brands now need source architecture, not click chasing.
CNN's Perplexity suit shows attribution is becoming an AI search product requirement, not a publisher courtesy.
Google's Preferred Sources and Highly Cited labels make source selection visible inside AI search.
Google AI Mode changes the visitor before the click. Sites now need proof paths, not just landing pages.
Google's new Search box turns queries into AI routing, making source architecture more important than rankings.
AI answer engines are changing links from destinations into evidence slots. Here is what source architecture now requires.
Google's I/O Search update moves discovery from ranked links to prompts, agents, and generated interfaces.
Google's 'disregard' failure shows why AI search needs source fidelity over answer polish.
Google's Gemini ad formats show AI search monetization moving from ranked results into synthesized answers.
Parallel Index points to the next AI search fight: pricing source value when agents consume the web.
Google's information agents shift AI search from one-off answers to persistent monitoring and source selection.
YouGov's AI Agent expansion shows AI search moving from public web results into governed private data.
GitHub's Copilot base-model change shows enterprise AI search is moving from optional model choice to governed defaults.
Google's spam policy now covers attempts to manipulate AI Overviews and AI Mode, raising the bar for AI search tactics.
Google's autocomplete AI Overview test turns query suggestions into answer prompts, changing where AI search begins.
Google's AI Mode link updates show AI search is becoming a source-selection interface, not just an answer box.