AI Overviews are eating your clicks. Why I'm not panicking — and what we do instead
Chandan KumarFounder · Performance Marketing Specialist
The numbers coming out of search right now are genuinely alarming if you only read one column of them. AI Overviews are showing on roughly 18% of Google searches and climbing. When that box appears, click-through to actual websites has fallen from around 15% to about 8% — close to a 47% collapse in clicks reaching publishers. Sixty percent of searches now end with no click at all. Gartner has been forecasting search volume dropping a quarter by 2026, and ChatGPT is busy answering the questions people used to type into Google. Every week a founder forwards me one of these stats with a worried note. And every week I tell them the same thing: read the other column before you panic.
Traffic is down — but you're watching the wrong number
Here's the column nobody screenshots. The traffic that still makes it through from AI search converts dramatically better — by some measures around 14% versus under 3% for traditional search, with roughly 42% higher conversion, more revenue per visit, lower bounce and longer time on site. Fewer people are arriving, and the ones who do are far closer to buying. That's not a catastrophe; it's a re-sort of who bothers to click.
Think about why. The casual searcher — 'what is X', 'how does Y work' — gets the answer in the AI box and never leaves. They were never going to buy from you this week anyway; they were tyre-kickers padding your analytics. The person who still clicks through has a real, specific need the summary couldn't satisfy. So when a founder shows me a scary drop in sessions, my first question is always: what happened to leads and revenue? Nine times out of ten they're flat or up. If you're still reporting pageviews to your board in 2026, you're measuring the era that just ended.
Win the citation, not just the ranking
The new game isn't only ranking on a page of blue links — it's being the source the AI quotes. If ChatGPT or Google's AI Overview names your brand in its answer, you get the trust and the mention even when there's no click, and you become the obvious destination when a click does happen. This is what people are clumsily calling AEO or GEO — optimising to be the answer, not just to appear somewhere near it.
Practically, that means content that genuinely, clearly answers real questions — structured, specific, and trustworthy enough for a model to lean on — plus the entity signals that make you a recognised brand: consistent information, schema, reviews and mentions across the web. For the Indian local businesses we work with, this overlaps heavily with Google Business Profile and review volume, because that's a lot of what the AI draws on when it recommends 'the best clinic in Indiranagar' or 'a reliable tour operator in Manali'. Being citable is the new being-rankable.
Paid search changes shape — defend and capture
For performance marketing specifically, the zero-click shift makes paid placement more important, not less. When organic clicks shrink and the AI box pushes the classic results further down the page, the paid slots that sit above or around that box become some of the only reliable ways to capture a high-intent searcher at the exact moment of need. The click is scarcer and, yes, often pricier — which changes the maths of what you do with it once you've got it.
It also makes branded-search defence non-negotiable: when someone has heard of you from an AI answer and then searches your name, you cannot afford to lose that click to a competitor's ad sitting on top of it. And because every paid click now costs more in a thinner market, the old full-funnel discipline matters more than ever. A scarcer, pricier click landing on a slow page with no follow-up is a far more expensive mistake in 2026 than it was in 2022. Fewer shots means each one has to be set up to convert.
The long game: be the brand the AI recommends
Step back and the real shift is this — AI search rewards brands, not pages. The model recommends names it 'trusts', and that trust is built the slow way: real reviews, genuine mentions, a consistent presence, customers who actually had a good experience and said so publicly. You cannot growth-hack your way into an AI's recommendation the way you could once hack your way up a results page with thin, keyword-stuffed pages.
That's why I keep telling founders this is a brand-and-performance moment, not a choice between them. The performance work captures and converts the demand that's there today; the brand work makes you the answer the AI offers tomorrow. The businesses that come out of this shift ahead aren't the ones chasing the lost clicks — they're the ones building the reputation that makes an AI hand them the customer. Fewer visitors, better visitors, and a brand the machine wants to recommend. Optimise for that, and the scary column stops mattering.
Key takeaways
- Stop reporting sessions; in the AI-search era, judge by leads and revenue from better-qualified visitors.
- Optimise to be cited (AEO/GEO) and defend branded and high-intent paid search aggressively.
- Fewer, costlier clicks reward a sharp offer, a fast page and instant follow-up — and a brand the AI trusts enough to recommend.
Written by
Mr. Chandan Kumar
Founder & Performance Marketing Director, Global Info Edge
Founder of Global Info Edge and a performance-marketing specialist with 15+ years in the digital marketing world — Google & Meta ads, conversion funnels and growth.
View full profile