Oct 30, 2026: a small study reported by Search Engine Roundtable’s coverage of AI Overviews/AI Mode CTR puts click-through from Google’s AI features at about 0.9%. The same write-up says sites that opted out of showing in Google’s generative AI saw fewer clicks from the regular web results too.
Our take: stop treating AI Overviews like “SEO traffic with a new label.” Treat them like brand visibility with near-zero sessions—and plan your month around winning the click you can still influence: classic organic listings, branded demand, and what happens after the user lands.
What actually changed (and what didn’t)
Nothing “shipped” in the sense of a new Google feature release. What changed is that we now have a concrete, if limited, benchmark that matches what many teams are feeling in analytics: AI Overviews can mention you and still send almost nobody.
The other important point in the report is the opt-out effect. If opting out reduces standard web clicks, it’s a warning sign for publishers who think they can simply block AI features and carry on as before. Even if the causality is messy (small sample, many variables), the business implication is simple: you may be trading a tiny slice of visibility for a meaningful slice of traffic.
Why 0.9% matters to business owners (forecasting is broken)
If you’ve been forecasting SEO upside assuming “rank #1 = X% CTR,” AI Overviews and AI Mode break that model. A citation inside an AI answer is not equivalent to position 1–3. In many cases it’s closer to being a footnote.
Here’s who wins and loses if this holds:
Winners: brands with direct demand (people searching for the brand name), brands that already convert well once visitors arrive, and businesses with defensible differentiation (price, inventory, local availability, guarantees) that doesn’t compress neatly into a generic summary.
Losers: publishers built on informational traffic arbitrage, “me too” service businesses relying on generic how-to pages, and any team still reporting success as “impressions went up” while pipeline is flat.
We usually recommend planning for AI Overviews traffic as bonus, not baseline. If you budget headcount or content production expecting AI citations to behave like blue-link clicks, you’ll overinvest in the wrong work.
Everyone will miss this: the opt-out story is the real threat
Most people will fixate on the 0.9% CTR. The more strategic problem is the reported drop in clicks to the main results after opting out. Even if it’s partly correlation, it highlights a nasty reality: Google can route attention across its own surfaces, and your visibility is not only a function of “ranking.”
So the decision isn’t “AI or not.” It’s “how do we stay eligible for visibility and make the visits we do get worth more?” That’s an SEO + conversion + brand question, not a robots.txt debate.
What we’d change on client sites this week (effort vs impact)
High impact, low effort (2–5 days): improve the snippet that competes for the click. Titles, meta descriptions, and on-SERP messaging are back to being money.
In practice, we’re rewriting templates for category/service pages so the first 60 characters carry one concrete differentiator (price bands, turnaround times, coverage area, inventory depth). This is not “SEO copy.” It’s ad copy for organic.
High impact, medium effort (1–3 weeks): build pages that close the loop after the AI answer. If AI Overviews reduce clicks, the clicks you keep must convert harder. That means faster pages, clearer offers, and fewer dead-end informational posts with no next step.
Medium impact, medium effort (2–4 weeks): strengthen branded demand. If generic discovery is getting absorbed by AI summaries, brand search becomes a safer channel. This is where digital PR and “why us” content matter again.
This is exactly where our search optimisation retainers are shifting: less “publish 10 articles,” more “win the click, then win the customer.” Pair that with landing page improvements through our web build team when the bottleneck is UX, speed, or lead capture.
What this means for your site
- Reset forecasts: treat AI Overviews and AI Mode mentions as awareness. Don’t put them in the revenue model unless you can measure downstream lift.
- Rewrite top templates: update titles and meta descriptions on your top 20 landing pages to compete aggressively for the remaining clicks.
- Upgrade “money” pages: add proof (case studies, pricing ranges, delivery times, comparisons) and remove friction from forms and checkout.
- Track the right thing: annotate when AI features spike for your head terms, then monitor changes in branded impressions/clicks in Search Console.
- Don’t panic-opt-out: test any AI opt-out decision on a small section first, because the reported side effect is lost web-result clicks.
Where Google is heading: training controls and eligibility splits
Part of the confusion is that “AI” now means multiple things: being eligible to appear as a cited source, allowing training, allowing grounding, and so on. Google is also moving toward simpler controls for publishers, which will make it easier for teams to choose their stance—without fully understanding the trade-offs.
If you want context, see Search Engine Roundtable’s report on Google’s new AI training control direction. Even if you never touch a control file, this is a sign that eligibility decisions will become more explicit—and more common.
Our opinionated stance
AI Overviews are overhyped as a traffic channel. They are not overhyped as a brand channel. If you’re in a competitive space (legal, home services, B2B software), “being the cited source” can still shape trust—especially for people who later search your brand name and convert.
But if you do nothing for 90 days, here’s what tends to happen: impressions look fine, rankings look “stable,” and sales quietly dip because the click share drains into AI answers and competitors with better snippets and stronger brands take what’s left.
If you need to tighten the whole funnel (technical, content, and conversion), start with a quick audit via our Tech and Search updates hub, then prioritise fixes on the pages that already have demand. If your product experience is the issue, our app delivery team can help remove the friction that makes fewer clicks hurt more.
Next step: pull Search Console for the last 28 days, list your top 20 pages by impressions, and rewrite those snippets before you publish another “AI-friendly” blog post.
Primary source: Search Engine Roundtable
