Are AI Overviews killing your traffic?

The AI Overviews traffic impact is real, but the panic around it is often bigger than the exposure most sites actually face. The largest independent look so far comes from the Pew Research Center, which tracked the real browsing activity of 900 U.S. adults and found that clicks roughly halve when an AI summary is present. That is a serious shift, but it is not the same as every page losing half its visitors. Whether AI Overviews dent your traffic depends almost entirely on the kinds of searches you rank for. This guide puts numbers to the impact, shows which query types trigger AI answers most, and gives you a way to measure your own exposure instead of reacting to headlines.

Are AI Overviews actually killing your traffic?

For informational and question-based content, yes, they are taking a real bite. For commercial and transactional pages, usually not much yet. AI Overviews replace the need to click by answering inside the results page, so the loss concentrates on the queries where a short summary satisfies the searcher.

The evidence is consistent across independent measurements. A randomized field experiment by researchers at the Indian School of Business and Carnegie Mellon University, reported by Search Engine Journal, found that AI Overviews reduced organic clicks by 38% on queries where they appeared, and pushed zero-click searches from 54% to 72%. That last number matters most: nearly three in four AI Overview searches now end without a click to any website. So the honest answer is that AI Overviews are not killing all traffic, they are killing the informational clicks you used to win from a featured snippet or a top-three ranking.

How much do AI Overviews actually reduce clicks?

Estimates range from roughly one-third to well over half, depending on the method and the query set. The spread is wide because a click-through model on high-visibility keywords will look worse than a broad browsing panel that includes low-competition searches.

Here is how the main independent measurements compare:

Source	Method	Reported click impact
Pew Research Center	Browsing panel, 900 U.S. adults	Link clicks fell from 15% to 8% of visits when an AI summary appeared
ISB / Carnegie Mellon field experiment	Randomized, 1,065 Chrome users	38% fewer organic clicks; zero-click rate up from 54% to 72%
Pew Research Center	Session behavior	Users ended their session 26% of the time with an AI summary, versus 16% without

The Pew session figure is the one publishers underrate. When an AI summary appears, a quarter of people simply stop browsing, because the answer was good enough. That is the mechanism behind the click loss, not a ranking penalty. Your page can still rank first and lose the visit before the choice to click ever happens.

Which searches trigger AI Overviews the most?

Long, question-shaped, informational queries trigger AI Overviews far more than short or commercial ones. Pew's data on trigger rates is the clearest public breakdown, and it is the practical key to estimating your own risk.

Query type	Share that produced an AI summary
One or two word searches	8%
All searches (baseline)	18%
Full sentences (noun plus verb)	36%
Searches starting with who, what, when, why	60%
Searches of 10 words or more	53%

Read that table against your own keyword list and the picture clarifies fast. If most of your rankings are short brand terms, product names, or buying queries, your exposure is low. If you built a blog on how-to guides, definitions, and "what is" explainers, a large share of your footprint sits directly in the blast radius. This is the information gain most panic-driven articles skip: exposure is a property of your query mix, not a flat percentage you apply to your whole site. For the flip side of the same coin, see how to rank inside Google AI Overviews so you at least earn the citation when one appears.

Why do some sites lose far more than others?

Because the loss follows content type, not domain size. Two sites with identical traffic can see wildly different damage if one lives on informational queries and the other on transactional ones.

Coverage from Forbes and other outlets points to the same pattern: how-to posts, explainer articles, definition pages, and comparison guides are hit hardest, while transactional pages hold up because a summary cannot complete a purchase. Sites that depend on top-of-funnel informational traffic to feed a funnel are the most exposed. Sites where visitors arrive ready to buy, book, or sign up feel far less, because those searchers still need to reach a real page to act. If your traffic model assumes a big informational top that funnels down, AI Overviews attack exactly that top layer. Understanding generative engine optimization becomes less optional the more your business leans on informational content.

How do you measure your own AI Overview exposure?

Use Google Search Console, not a scare stat. Your own impression and click data will tell you exactly which queries are bleeding, and you can do it in an afternoon without a paid tool.

A simple, repeatable method:

• Open the Performance report and compare a recent period against the same period last year, on the same season.
• Sort queries by clicks lost. The queries where impressions held steady but clicks dropped are your AI Overview casualties.
• Tag those queries by shape. Question words, "how to", "what is", and 10-plus word phrases are the high-trigger ones from the Pew table above.
• Estimate exposure as the share of your total clicks that came from those high-trigger queries. That percentage, not a headline number, is your real risk.

This turns a vague fear into a portfolio view. You will usually find that a handful of informational pages account for most of the loss, while your money pages are untouched. That is a very different, and far more actionable, situation than "AI is taking half my traffic." Our measurement methodology walks through the reporting setup in more detail.

Does the traffic that survives convert better?

Often, yes. When a summary answers the casual, low-intent searchers, the people who still click through tend to be the ones who genuinely need more than a summary can give. Several analyses report that surviving click-through traffic engages more deeply and includes more qualified leads, because the tire-kickers were filtered out by the answer box.

This does not make the click loss painless, especially for ad-funded publishers who monetize raw pageviews. But for a business selling a product or service, a smaller pool of higher-intent visitors can be worth more than a large pool of definition-seekers who were never going to convert. The strategic response is to stop counting all clicks as equal and start valuing intent. That reframing matters because it changes what you should optimize for next: being the source the AI cites and recommends, not just the page that ranks.

How do you recover traffic lost to AI Overviews?

You do not win it back by ranking higher on the same page. You win by getting cited inside the AI answer and by earning visits from surfaces Google does not control. The recovery playbook has three parts.

1. Get cited inside the Overview. Structure content so Google's system can lift a clean, self-contained answer from your page and attribute it to you. Answer-first passages, clear definitions, and sourced statistics are what get pulled. A citation keeps your brand visible even on a zero-click search, and it is the closest thing to reclaiming the position.

2. Diversify beyond Google. Referrals from the AI engines, ChatGPT, Perplexity, Gemini, and Claude, are a growing traffic source that behaves differently from Search. Building visibility there hedges against any single answer box. If competitors keep showing up in those answers and you do not, our guide on why AI recommends competitors instead of you covers the specific gaps to close.

3. Own an audience. Email lists, communities, and return visits are traffic no answer box can intercept. The publishers coping best are the ones building direct relationships rather than renting attention from a results page.

Should you panic, or adapt?

Adapt. As a Northwestern Kellogg Insight analysis frames it, the shift rewards the sites that evolve their model rather than mourn the old one. AI Overviews are a permanent feature of search now, so the durable move is to measure your real exposure, protect your money pages, and rebuild informational visibility as citations and audience rather than raw clicks.

If you want to systematize the citation and tracking side, some platforms now bundle answer-first content structuring with AI-visibility monitoring across the engines. Sorank, the tool built by this site's operator, is one such platform and offers a 3-day free trial from $99/mo (disclosure: we operate it, and we cover competing tools on the same terms). Whatever tool you use, the strategy is the same: stop optimizing only for the blue link, and start optimizing to be the answer. Compare your options in our buyers guide before committing to anything.

Conclusion

The AI Overviews traffic impact is real but uneven: clicks roughly halve when a summary appears, yet the damage concentrates on informational, question-style queries and largely spares commercial and transactional pages. Measure your exposure in Search Console before you react, protect the pages that convert, and rebuild informational visibility as citations and owned audience. The sites that treat this as a shift to adapt to, not a disaster to survive, are the ones that come out ahead.

Compare the best AI SEO software

Frequently asked questions

Sources

• Pew Research Center: Google users are less likely to click when an AI summary appears
• Search Engine Journal: Study confirms AI Overviews cut organic clicks 38%
• Forbes: Google AI Overviews are eating your website traffic
• Kellogg Insight (Northwestern): As AI eats web traffic, don't panic, evolve
• Search Engine Land: New data on AI Overviews and click-through rates