Numéro · 7 mai 2026
Google AI Mode Just Hit 75 Million Daily Users. The Marketing Stack Has Not Caught Up.
Google AI Mode Just Hit 75 Million Daily Users. The Marketing Stack Has Not Caught Up.
The week the AI search story crossed the line from "interesting trend" to "your top-of-funnel just changed." Google quietly expanded AI Mode and AI Overviews on May 6 with article suggestions, in-response links, news-subscription highlighting, and previews of public-discussion perspectives. AI Mode is now processing over 1 billion queries a month with 75 million daily active users, the fastest adoption curve of any new search feature Google has ever shipped. Independent measurement from Seer Interactive: 93% of AI Mode queries get zero clicks, organic CTR has dropped 61% on queries with AI features, and sites cited in AI Overviews see a 35% click lift vs non-cited top-10 results.
Translation, in plain language. The traffic you used to win by ranking is now being absorbed at the top of the page. The traffic you do win is more valuable per visit, because the AI pre-qualified the user before sending them. Citation has replaced position. Most marketing dashboards still measure position. That mismatch is the entire story this week, and the bottom line for any CMO is that the SEO stack you renewed in January is already lying to you.
TL;DR
Google AI Mode is at 1 billion monthly queries and 75 million daily active users, and 93% of those queries result in zero clicks. Citation, not ranking, is now the primary AI distribution metric.
5WPR's AI Platform Citation Source Index analysed 680 million citations across the five major engines and found Reddit cited at roughly 40% frequency across all of them. YouTube has a 200x citation advantage over every other video source.
Search Engine Journal's Dan Taylor argues that scaled AI content fails because of weak content strategy, not because of the AI tools. Google's quality threshold is the filter most teams underestimate.
Jellyfish launched Share of Model, a product that turns AI-assistant brand mentions into Google Performance Max targeting signals. Brands are no longer just optimising for AI mentions, they are converting them into paid-media inputs.
Adthena's AdBridge converts Google Ads campaigns into ChatGPT-ready campaigns with keywords and competitive insights. The infrastructure for buying ads inside answer engines is being assembled in public.
1. Google AI Mode Crosses 75 Million Daily Active Users, Adds Community Perspectives
May 6, 2026 | Google announced new updates to AI Mode and AI Overviews to help users find relevant websites, deep insights, and original content from across the web. Users now see article suggestions, direct links within responses, previews of websites and personal perspectives, and a new feature that highlights links from their existing news subscriptions. AI responses now include previews of perspectives from public online discussions and social media. As of early 2026, AI Mode processes over 1 billion queries a month with 75 million daily active users. Seer Interactive's measurement: 93% of AI Mode queries get zero clicks; for queries with AI features, organic CTR dropped 61% from 1.76% to 0.61%; sites cited in AI Overviews see a 35% click lift vs non-cited top-10 results.
Source: https://blog.google/products-and-platforms/products/search/explore-web-generative-ai-search/
What I Think: Google added "perspectives" because the AI summary alone is not enough to keep users on Google. They are adding human-source previews to retain the click. The honest read: AI Mode at 75 million daily is a more aggressive adoption curve than mobile search and faster than voice search ever managed. Anyone arguing AI search is "still early" is using last year's data. The shift is happening at consumer scale right now.
Why It Matters: If your marketing dashboard reports rank position as a primary KPI, you are measuring the wrong thing. Switch to citation tracking immediately. The 35% click lift on cited results is not a nice-to-have, it is the entire upside available in AI Mode traffic. Either your content gets cited, or your category goes silent.
2. The 680 Million Citation Map: Reddit at 40%, YouTube at 200x, Wikipedia Everywhere
May 1, 2026 | 5W Public Relations released the AI Platform Citation Source Index 2026, the first consolidated ranking of the 50 websites most cited by generative AI answer engines. The analysis synthesises over 680 million individual citations across ChatGPT, Google AI Overviews, Perplexity, Gemini, and Claude, drawn from six major published citation studies between August 2024 and April 2026. Reddit is cited at roughly 40% frequency across every major LLM. YouTube holds a 200x citation advantage over every other video source and dominates Google AI Overviews. ChatGPT concentrates on Wikipedia, Reddit, Forbes, and Business Insider. Perplexity rewards primary sources, NIH/PubMed, and named B2B authority sites.
What I Think: This index is the closest thing the industry has to a single map of who actually decides what AI says about your brand, and the map is brutal. Most of the surfaces are not your owned media. Reddit, YouTube, Wikipedia, NIH, Forbes, Business Insider. The brand site is, at best, a co-equal source. Often a tertiary one. Pretending otherwise is the most expensive marketing assumption I see today.
Why It Matters: Audit your citation footprint engine-by-engine. The same brand can be cited 30% of the time on ChatGPT and 4% on Perplexity because the engines pull from different source sets. Build a content strategy that places primary sources on the platforms each engine actually rewards. For most B2B, that means LinkedIn long-form by named experts, peer-reviewed citations, and a clean Wikipedia entry. For most consumer brands, that means Reddit, YouTube, and a credible third-party reviewer.
3. Google's Quality Threshold Is Quietly Killing Scaled AI Content
May 7, 2026 | Dan Taylor at Search Engine Journal argues that the recent collapse in scaled AI-content traffic is not a Google penalty against AI itself, it is the Quality Threshold doing what it was always designed to do. Initial freshness boosts produce a traffic surge for new sites and pages. After a few months, Google's quality systems re-evaluate the content against engagement signals, dwell time, repeat visits, and editorial coherence. Pages that were generated faster than they were edited fall off. The article frames this as a strategy problem, not a tooling problem.
What I Think: Every AI-content vendor in the last 18 months has been selling speed. The data Dan is referencing is the post-honeymoon collapse, and it is consistent across every site I have looked at. Volume without editorial rigour produces a 90-day spike followed by a 12-month decline. The question is no longer "can you produce 1,000 articles a month." The question is "can you make any single article worth coming back to." Those are different muscles.
Why It Matters: Audit your content output from the last 12 months. Cut the 30% of pages with the lowest dwell time and lowest return-visit rate. Replace volume targets with engagement targets. The teams that survive 2026 will be the ones that stopped publishing 80% of what their AI tooling could produce, not the ones who produced more.
4. Jellyfish's Share of Model Turns AI Mentions Into Performance Max Targeting Signals
May 4, 2026 | Adweek reported that Brandtech-owned Jellyfish is pitching Share of Model, a product that compares how often ChatGPT, Gemini, and other AI assistants mention a brand versus its competitors, analyses themes and sentiment, and feeds those insights as targeting signals into Google Performance Max campaigns. The pitch closes a loop most marketers had been treating as separate: AI mention monitoring on one side, paid-media activation on the other.
What I Think: This is the first product I have seen that treats AI brand monitoring as an upstream input to media buying rather than as a standalone analytics output. The conceptual shift matters. If a brand is mentioned twice as often as a competitor inside ChatGPT recommendations, that is a real signal about market perception, and routing it into Performance Max gives the algorithm a richer fitness function. The strong version of this idea ends with a CMO who buys ads based on AI conversational share, not just keyword search volume.
Why It Matters: Even if you do not buy Jellyfish's product, the structure they are pitching is the right structure. Set up monthly AI-mention tracking against three named competitors. Feed the gaps into your media plan. The signal is too cheap and too predictive to ignore.
5. Adthena Builds a Bridge Between Google Ads and ChatGPT Ad Buys
May 1, 2026 | MarketingProfs covered Adthena's AdBridge, a tool that converts Google Ads campaign data, keywords, audience segments, and competitive insights, into ChatGPT-ready campaigns. The product is positioned as the migration layer for the new ChatGPT advertising surface that OpenAI has been gradually opening to brands. Adthena is one of several martech companies racing to build the infrastructure that will let brands buy answer-engine ad placements with the same workflow they already use for search.
What I Think: AdBridge tells you the boring part of the AI search transition is happening. The plumbing is being laid. By the end of 2026, your media agency will treat ChatGPT ads as a line item alongside Google Search and Performance Max. The brands that wait for the official OpenAI announcement will arrive 18 months late, with worse data and higher CPMs. The brands paying attention now are quietly testing AdBridge-style migrations and building the data hygiene that those campaigns need.
Why It Matters: Audit your Google Ads keyword set for export-readiness. Strip out the ones that depend on syntax tricks and only work because of Google's quality score. The keywords that travel cleanly into ChatGPT are the ones that match natural-language conversational intent. That is a different keyword discipline. Start re-writing the ad library now, before your agency has to do it under deadline pressure.
Travel Desk
For travel brands specifically, this week's GeoTravel data made the broader story concrete. Booking.com showed up in 95.3% of the 450 hotel queries Nokumo tested across four AI models. Hotel chain websites contributed only 4.3% of all cited URLs. Independent hotels 11.8%. NerdWallet was cited 13.6% of the time on Hyatt-specific queries, ahead of Hyatt's own site at 10.3%. The travel brand site is now a co-equal or tertiary source on questions about its own properties. The action items for travel are different from the cross-industry ones above: audit the OTA listing as if it were the brand site, fix the Wikidata Q-item (4.8x lift in AI Overview citation per Ahrefs' analysis of 4 million URLs), and place a 40 to 60 word plain-text answer capsule at the top of every page section. Full travel breakdown in this week's GeoTravel newsletter at Geotravel.ai.
The Strategic Reality Check
The single thread running across this week's news is that AI search is no longer a parallel channel that marketers can defer. It is replacing the top of the funnel for an audience that already numbers 75 million daily active users on AI Mode alone. The 5WPR map of 680 million citations confirms that the surfaces that decide brand visibility inside answer engines are mostly not the brand's owned properties. The Search Engine Journal piece confirms that the post-honeymoon traffic collapse is real and consistent. The Adweek and MarketingProfs coverage confirms that the paid-media infrastructure is being assembled in the open, and that the brands paying attention are already testing it.
Three truths that should focus the next 30 days for any CMO regardless of vertical:
(1) Citation has replaced ranking as the primary distribution KPI. Switch your dashboards before next month's QBR, not after.
(2) The surfaces that drive AI citation are mostly third-party. Reddit, YouTube, Wikipedia, NIH, named expert posts on LinkedIn, third-party reviews, the relevant industry trade publication. Build a programme that places primary sources on those surfaces, not just on your owned blog.
(3) The paid-media transition is happening underneath the discovery transition. Brands that build the AdBridge-style migration now will pay 30 to 50% less per acquisition next year, because they will have the data hygiene the new platforms require. Brands that wait will pay agency premiums to clean up the keyword set on someone else's deadline.
What You Should Do This Week
Set up citation tracking for ChatGPT, Perplexity, Gemini, AI Mode, and Bing Copilot against three named competitors. Free tools exist (Otterly, AmiCited, Bing Webmaster Tools AI Performance Report). Pick one and run a baseline scan today. Capture the result. That is your benchmark for the next 90 days.
Audit your last 12 months of content output for the post-honeymoon decline pattern Dan Taylor described. Pull the engagement and return-visit metrics for every page published since May 2025. Cut the bottom 20% by dwell time. Re-publish the next 20% with stronger answer capsules and original data. The math is simple. Quality content compounds, low-quality content decays, and the cliff arrives at the 90 to 120 day mark.
Build the Wikipedia or Wikidata entity record for your brand and your top three product lines. Both ChatGPT and Gemini lean heavily on Wikipedia and Wikidata for entity grounding. The work takes a few hours and the lift compounds across every AI engine that uses entity recognition. The cash cost is zero. The risk is zero.
Brief your media agency on AdBridge-style migration. Even if you are not buying ChatGPT ads yet, the prep work, conversational keywords, audience segments aligned to AI surfaces, competitive intelligence that survives platform changes, is overdue. Ask your agency for a written 90-day plan. If they cannot produce one, you have a different problem than AI search.
Stop measuring SEO success with rank-tracking software alone. Add a citation share metric to your monthly reporting. Define a target. Brands that report citation share alongside rank position are 18 months ahead of the brands that do not.
Think your category is too niche or too B2B for AI search to matter? 5WPR's index shows Perplexity rewards NIH/PubMed and named B2B authority sites at higher frequency than ChatGPT does. The B2B AI search shift is happening at every level of seniority. Hit reply with your category and I will tell you which engine indexes it most aggressively.
Want a 30-minute AI visibility audit before your competitors run one on you? Let's talk: https://calendly.com/vincent-getinference/
See you next week.
Vincent Chief AI Enthusiast, SimplyAI: AI SEO (GEO)