AI Overviews and ecommerce Google Ads: what actually changes

Here is an uncomfortable pattern we see across ecom accounts and our own site: rankings hold, impressions climb, clicks fall. Positions that used to mean traffic don't anymore. The page is right where it always was. The buyer just never scrolls down to it, because Google already answered the question at the top of the page.

That is AI Overviews doing exactly what they were built to do. And if you run an ecom brand on Google Ads, you need to know what this changes, what it doesn't, and what to do about both.

What AI Overviews actually are

The key word is some. Google decides per query whether an AI Overview shows. Ask "what is a good ROAS for jewelry" and you will probably get one. Search "gold hoop earrings 14k" and you will mostly get Shopping ads, product grids, and stores - the stuff that makes Google money.

That split matters more than anything else in this post. AI Overviews eat the questions. They have not eaten the buying.

How AI Overviews affect ecommerce traffic

Think about the two kinds of searches your customers make.

Research searches. "How to layer necklaces." "What size rug for a living room." "Are linen sheets worth it." These used to send readers to blogs and guides, where good stores warmed them up and cookied them. Now Google answers a big share of these at the top of the page. The user reads, nods, and leaves. Your guide still ranks. It just doesn't get the click it used to earn.

Buying searches. "Linen sheets king size white." "Buy 14k gold hoops." These still behave mostly like they always did. High intent, product grids, Shopping carousels, stores. This is the ground Google Ads was built to win, and it is largely intact.

So the honest summary: AI Overviews are an organic-traffic problem before they are a paid-traffic problem. If your growth depended on blog clicks from informational keywords, you are feeling this already. If your growth runs on Shopping and Performance Max, the direct hit is much smaller - for now.

Why your Shopping ads are mostly fine

Google is not going to starve its own revenue. Shopping placements, Performance Max, and search ads still render on commercial queries, including alongside AI-shaped results. When someone shows buying intent, Google still wants to show them products, because that is the auction that pays for everything else.

Two things are worth watching, though.

First, layouts keep moving. Where ads sit relative to an AI answer changes between markets and query types, and Google tests constantly. Don't build your expectations on a screenshot someone posted from another country. Watch your own impression and click-through data.

Second, the query mix underneath your campaigns is shifting. People are learning to search the way they talk to a chatbot - longer, more specific, more conversational. That changes which search terms trigger your ads. Which brings us to the paid-side playbook.

What to do on the paid side

Protect your brand terms. When someone searches your brand name, you want your ad and your site at the top, full stop. An AI Overview summarising "is [your brand] legit" from third-party sources is not a summary you control. A cheap brand campaign guarantees you own that moment. This was smart before AI Overviews. It is non-negotiable now.

Live in your search term reports. The queries feeding your campaigns are changing shape. Longer questions, comparison phrasing, "for" and "vs" and "which" constructions. Some of these are gold - a long conversational query often carries very clear intent. Some are junk you should negate. Either way, you only see the shift if you look. Make search term review a weekly habit, not a quarterly cleanup.

Feed quality just got a second job. Your product feed already decides how well Shopping performs. Now the same structured data feeds whatever AI-shaped shopping experience Google builds next. Clean titles, correct attributes, honest pricing - that work now pays twice. We wrote up how to optimize your product feed for AI shopping as its own piece, because it deserves one.

Don't panic-cut informational awareness spend based on vibes. If a campaign's job was catching researchers early, check whether its numbers actually moved before you kill it. Decisions on data, not on headlines.

The new organic play: be the source the AI cites

Here is the part most stores miss. AI Overviews don't invent answers. They are built from pages - and they cite some of them. Those citation links are the new organic click.

Fewer clicks get distributed, but they go to the sources the AI trusted enough to quote. That flips the content goal. The old game was ranking position. The new game is being extractable enough that the machine picks you as its source.

What extractable looks like in practice:

Answer first. Put a direct, standalone answer in the first lines under a heading that matches the question. Two or three sentences that hold up if lifted out with zero context. Then go deep for the humans who keep reading.

Real tables. A comparison table with actual values beats four paragraphs of prose describing the same thing. Machines lift tables cleanly. So do skimming humans.

Specific claims. "Founded 2024, 200+ ecom brands, €200M+ in client revenue generated" is quotable. "A leading agency with proven results" is fog. AI systems quote facts, not fog.

One page, one question. A page trying to answer twelve questions gets cited for none of them. Tight pages that fully own one query win extraction.

This is the same preparation that matters for conversational platforms generally - we covered the ad side of that shift in ChatGPT ads vs Google and Meta. Different surface, same fundamentals: structured, specific, trustworthy content wins.

What nobody knows yet

Time for the honest section.

Nobody outside Google knows exactly which queries trigger AI Overviews next quarter, or how the click-through math settles once user behaviour stabilises. Nobody knows what an "ad inside an AI answer" auction looks like at scale, though everyone expects one. Anyone selling you a guaranteed AI Overview placement strategy is selling confidence, not knowledge.

What we do know: informational organic clicks are shrinking, buying-intent placements still work, and cited sources still get traffic. That is enough to act on. Build for those three facts, re-measure monthly, and adapt when the ground moves - because it will. The people watching their own data will adjust in weeks. The people waiting for certainty will adjust years late.

And if you are wondering whether the humans running all this are about to be automated away too, we gave that question an honest answer of its own.

Frequently asked questions

Do AI Overviews hurt ecommerce stores?

They hurt one specific thing - organic clicks on informational queries. When Google answers "how to choose a standing desk" at the top of the page, fewer people scroll to the blog posts below it. Rankings can hold steady while clicks fall. High-intent product searches are less affected, because someone ready to buy still needs a store, and Shopping placements still render on those queries.

Do Shopping ads still show when an AI Overview appears?

Yes. Shopping and Performance Max placements still render on commercial queries, including around and inside AI-shaped results. Google makes money from those placements, so it has every reason to keep them visible. The exact layout keeps changing, which is why you watch your own impression and click data instead of trusting screenshots from other people's markets.

How do I get cited inside an AI Overview?

Make your content easy to lift. Put a direct, standalone answer at the top of the page, right under the heading that matches the question. Use clean comparison tables with real values instead of prose walls. Keep claims specific and sourced. AI Overviews are built from pages the system can extract cleanly, so the pages that answer first and structure their facts get quoted.

Should I change my Google Ads setup because of AI Overviews?

Not drastically. Keep brand terms covered so an AI answer never becomes the only thing standing between a branded search and your site. Read search term reports more often, because the query mix shifts as people ask longer, more conversational questions. And keep your feed clean - the same product data that feeds Shopping is what AI surfaces read.