While everyone was busy mastering AI prompts, something quieter happened to organic search traffic. It started disappearing before our eyes.
Google keeps saying it sends more traffic than ever, but publishers from the BBC to Bloomberg report the opposite. Gartner has predicted organic search volume could drop around 25% as AI answers absorb the clicks. My own bet sits closer to 20%. And here’s the twist: Google searches are still growing. The queries keep climbing. Google is just keeping the traffic for itself.
Let’s get into what’s actually happening, and what I’d do about it.
The zero-click reality, by the numbers
Treat these figures as directional, not gospel. They shift quarter to quarter as the labs leapfrog each other. But the direction has been clear for a while.
Roughly 6 in 10 Google searches now end without a click to any outside website. Let that sit for a second. The majority of searchers never leave Google. SparkToro’s click-stream research has been making this point for years, and “zero-click” is now the default experience, not the edge case.
A couple of related patterns from that same body of research:
- For every 1,000 U.S. searches, only a few hundred clicks reach the open web. The rest stays on the results page or routes to a Google property.
- A meaningful share of the clicks that do happen land on another Google property: YouTube, Maps, Flights, Shopping. That shrinks the external-traffic pie further.
This isn’t a forecast anymore. It’s the baseline you’re already operating in.
The AI Overviews effect made it worse
If the zero-click trend was the slow leak, AI Overviews (and the newer AI Mode in Google Search) turned it into a faster one.
When an AI Overview shows up above the results, the top organic listing’s click-through rate takes a real hit. Seer Interactive’s ongoing CTR tracking found organic click-through on informational queries falling sharply once AI answers entered the page, in some measurements by more than half. The exact percentages bounce around depending on the query set and the month. The pattern doesn’t.
So here’s the takeaway. If your traffic leans on informational queries (the “what is,” “how does,” “best way to” terms), AI Overviews are quietly answering those for you, on Google, without the click.
This is the core of what generative engine optimization is built to address. The game is no longer only “rank #1.” It’s “be the source the AI answer pulls from, and own the demand that survives the zero-click page.”
Five moves I’d make this quarter
1. Shift toward lower-funnel keywords
Transactional and comparison queries still earn healthy clicks, because the searcher needs to leave and do something: buy, sign up, compare, book.
Do this: Audit your keyword portfolio. Flag informational terms that now trigger an AI Overview and reallocate effort toward purchase-intent terms. Test adding modifiers like “buy,” “pricing,” “vs,” and “alternatives” to terms you already target.
2. Treat zero-click visibility as a real outcome
Being cited inside Google’s AI answer delivers awareness even with no visit. That’s uncomfortable for traditional traffic reporting. It’s also the reality.
Do this: Track when and how your brand shows up inside AI Overviews and AI answers. Most rank-tracking tools (Semrush, Ahrefs, and others) now flag AI Overview presence. Watch it like you watch rankings.
3. Build content AI can’t fully swallow
Some formats keep earning clicks despite the summary sitting above them. AI is good at compressing “what.” It’s weak on “how, specifically, in your exact situation.”
Do this: Lean into original frameworks, your own data, interactive tools, and walkthroughs grounded in real work. The litmus test: does the piece answer “what” (easy for AI to absorb and replace) or “how I’d actually do it” (worth the click)?
4. Diversify your traffic sources now, not later
A rough allocation I keep coming back to:
- ~50% direct and owned channels (email, SMS, community, podcast)
- ~30% search (paid, traditional organic, Bing)
- ~20% video (YouTube, connected TV, marketplace search)
The shift: Create for your direct subscribers first, optimize for search second. The newsletter you own can’t be zero-clicked away.
5. Make your pages AI-friendly with structured data
You can’t force a citation. You can make yourself easy to cite.
Do this: Implement FAQ, HowTo, and table structured data on high-value pages. Put a clear, concise definition near the top of each article so an AI answer can lift it cleanly. And check that your IT team didn’t quietly block the AI crawlers (GPTBot, ClaudeBot, PerplexityBot) you actually want pulling your content.
New KPIs for an AI-search world
Traditional SEO metrics were built for a world where ranking meant traffic. That link is fraying. Here’s where I’d start, then adapt for your industry.
Zero-click and citation metrics
- AI answer capture rate: how often your content appears inside AI Overviews and AI answers
- Knowledge panel and answer-box presence for your brand and core terms
Search intent quality
- Topic authority: how well your content reads as authoritative to an AI, tested by running your top 20 industry questions through ChatGPT, Gemini, Claude, and Perplexity and seeing who gets cited
- Query satisfaction: are you the answer, or a stepping stone to someone else’s video?
- Depth before breadth: go deep on a tight topic set first, then expand
Multimodal and technical readiness
- Visual and voice search visibility for “near me” and conversational queries
- Cross-format citation: how often AI references you across text, video, and image
- Structured data completeness plus Core Web Vitals (and confirming you didn’t block the AI crawlers)
Brand-specific AI metrics
- Brand citation frequency: manually query how often each AI mentions you for your top 20 relevant prompts
- Share of voice vs. competitors inside AI results
- Sentiment: how does the AI actually frame your brand when it brings you up?
- Referrer segmentation: in GA4, segment traffic from known AI referrers (ChatGPT, Perplexity, and the like) so the assist traffic doesn’t hide in “direct”
If model routing is part of how your team works now (and it should be), this connects to the wider AI marketing stack: the same tools answering search queries are the ones you’re using to produce, test, and measure the content.
A quick win for this week
Pick three of those metrics and stand up a simple “AI search dashboard.” Then run your top 20 target keywords through both traditional search and the AI answer interfaces, and note the gaps. Where does the AI answer the question completely? Where does it cite a competitor and not you? Where does it cite nobody, leaving an opening?
That gap list is your roadmap.
Where this goes from here
SEO isn’t dead. It’s molting. The teams adapting fastest are reallocating effort toward bottom-funnel content Google is less likely to cannibalize, and toward owned channels that can’t be summarized out from under them. Treating on-SERP visibility as a goal feels strange if you grew up reporting sessions and clicks. It’s also the reality we’re in.
One more move worth the five minutes: add a “How did you hear about us?” field to your lead and post-purchase forms, with “ChatGPT or another AI assistant” as an option. Most AI-influenced buyers won’t click a trackable link. They’ll read about you in an answer, then show up later by typing your name. That survey field is one of the few ways to catch the demand the analytics will never show you.
Then ask the question I keep asking myself. If I got zero clicks from my top informational queries next month, how would I replace that traffic? If you don’t have an answer yet, that’s the work.
Get the zero-click playbook before the next algorithm shift
I spend a lot of my week watching what actually survives the zero-click page: which pages still earn the click, which ones get cited inside the AI answer and pull in buyers anyway, and which old SEO moves quietly stopped working this month. I write down what I find so you don’t have to lose the traffic first to learn it.
Subscribe free → and adapt before the traffic does.