Google now has a billion people a month using a version of search that answers the question for them instead of handing over ten links. Sundar Pichai put the number on the I/O stage in May: 1 billion monthly actives for AI Mode and 2.5 billion for AI Overviews, with AI Mode queries running three times longer than classic searches. Eight weeks on, I think most of the hot takes still miss, some by overreacting and some by waving the number away.
So this is my version for founders. I'll stick to the numbers that matter, the paradox hiding inside that billion, and what I'd change this quarter.
The number, and what it actually counts
Nobody needs to doubt the headline figure. It came straight from Pichai on stage at I/O in May 2026, and Google's own recap puts AI Mode past 1 billion monthly active users. That's the conversational tab, a different thing from the AI Overviews summary box, and it got there about a year after launch, one of the fastest ramps Google has ever had. It now runs in nearly 200 countries and 98 languages, with Gemini 3.5 Flash as the default model.
The asterisk is what should shape your reaction. Monthly actives count anyone who touched the product, and Google has been pushing AI Mode in front of people hard. Third-party desktop data that's been going around the SEO community puts AI Mode at less than a quarter of one percent of actual search events. A billion people touch it, and very few of them live in it. The sharpest read of that gap I've seen came from a tech analyst:
I'm not sure if Google is winning the AI race. However, I think they're winning the AI distribution race, which is a different thing. 900M Gemini users is impressive on a slide. But a huge chunk of that is Android users who got a default app swap and Search users who got AI https://t.co/WXZxznAUxh
Chubby♨️@kimmonismusMay 26, 2026
Paraphrasing his argument: Google may or may not be winning the AI race, but it is clearly winning the race to distribute AI, because a huge share of those users got AI surfaces by default and never chose them. For visibility planning, distribution is the race I care about. Adoption that was pushed on people still changes where your buyers' answers come from.
The skeptics keep ending up in the same place:
google processes 3 quadrillion tokens a month and its AI mode search (which isn't even that good!) has a billion users, i think they'll be ok
Kevin Roose@kevinrooseMay 29, 2026Whatever you think of the quality, a surface with a billion users is too big to ignore, even when the NYT's tech columnist rates the product "not even that good."
What it's doing to traffic
I don't think SaaS founders are the ones this traffic collapse is burying, but the numbers are rough enough that you should see them. SparkToro's 2026 clickstream analysis found that 68% of Google searches now end without a click to the open web, against 60% in 2024. Eight points in two years is a steep climb for a number that was already bad.
The pain isn't spread evenly. Chartbeat has two years of tracking on more than 2,500 content sites, widely shared after I/O, and it shows small sites down roughly 60% in Google traffic and mid-sized ones down 47%, with even sites doing 100K daily visits down 22%. Publisher trade data agrees. In the biggest thread on the subject the mood is openly funereal, and publishers are talking about opting out of Google entirely:
Once Unimaginable, Publishers Are Preparing to Opt Out of Google Search
The most upvoted take there, paraphrased, is that Google spent twenty years getting the web to depend on search traffic and then turned search into a machine that answers without sending any. A B2B commenter in the same thread reported double-digit traffic drops across professional services, with AI sending no clicks to replace them.
That funeral is for one business model, the one built on ad-monetized pageviews, and SaaS founders aren't part of it. Your economics depend on being the product a buyer puts on the shortlist, and there's counter-evidence that sitting inside the answer pays: one study found brands mentioned in AI summaries picked up meaningfully more clicks on the same queries. The game is still on, it's just played in the answer box now, which is the shift we laid out in the zero-click playbook.
Know exactly what AI says about your competitors.
RankControl's Recon Agent monitors competitor citations across ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode. See where they show up and you don't.
Your rank tracker is watching the wrong screen
The reason your dashboards look calm through all of this is mechanical. AI Mode doesn't run the query you typed. It breaks it into 10-15 sub-queries and pulls candidates for each one before writing an answer with citations, so the pool of cited pages ends up looking very little like the SERP you track.
One analysis of AI Mode citations found that the overwhelming majority weren't in the organic top 10 at all. If positions one through ten are all you watch, most of what AI Mode cites never crosses your screen, and BrightEdge's long-running overlap data shows AI citations and top-10 rankings staying stubbornly apart.
Google's answer to "how do I see this?" arrived on June 3, when Search Console got a generative AI performance report showing impressions of your URLs inside AI Overviews and AI Mode. You get impressions and that's it, with no clicks and no prompts. The data only goes back to May 18, and the report is still rolling out. Practitioners worked out what that means pretty fast:
How to know for which search terms we are getting sessions from LLMs?
I want to track the search terms from ChatGPT, Gemini, Claude etc., just like Google Search console, can anyone help me with this.Thanks in advance.
Their consensus, paraphrased: quit trying to reconstruct queries you'll never see, and instrument what you can observe, which is who gets cited for the prompts that matter. That idea is the whole design behind our AI visibility tracking. We sample every customer's buyer prompts across AI Mode and the other engines and log citation share per query, because that number moves before any traffic chart does.
26 content formats. Published on your domain. Matched to your brand.
Guides, comparisons, listicles, case studies, and more. RankControl generates content that gets cited by ChatGPT, Perplexity, Claude, Gemini, Grok, and Google AI Mode.
What to change this quarter
I'd make four changes, listed here roughly by payoff. Start measuring share-of-answer next to rank: pick your 20-30 buyer prompts, sample them every week across AI Mode and the chat engines, and log who gets cited. Rank still matters as an input, but the answer is the output. Share of voice is the metric now.
Then cover the fan-out. Take each head term you care about, list the sub-questions AI Mode would split it into (pricing, comparisons, integrations, "is it worth it"), and make sure every one has a self-contained, answer-first passage somewhere on your site. Question-shaped H2s followed by one-line answers extract best.
The third change is about the surfaces AI Mode already trusts: YouTube, Reddit threads, review platforms and comparison listicles. On B2B software queries, community and review pages regularly out-cite vendor sites, which is why an updated G2 profile can beat yet another landing page.
The last one is an audience lever. Google's Preferred Sources now works inside AI Mode, and users who select a source are twice as likely to click it. A newsletter or community that gets buyers to pick you was already a retention play, and now it's an AEO tactic as well.
I'd bet the number is bigger next quarter. Ads already show up in a quarter of AI Mode results, and agentic search is rolling out. Every earnings call will bring another adoption stat on top of that, and none of it should change your plan, because the plan was never to wait and see.
It was to instrument the answers and cover the questions, and showing up on the surfaces those answers trust is part of the same job. The blue links won't die this quarter, but the decision has already moved out of them. We laid the groundwork in our first AI Mode guide, and the billion-user milestone took the "maybe this won't matter" scenario off the table.
Your competitors are getting cited by AI. You're not.
Every day without citation tracking is a day your competitors pull ahead in ChatGPT, Perplexity, and Claude.
