AI Overviews and your Shopify traffic: what to expect

The short answer

Expect meaningfully fewer clicks on the informational queries an overview can answer, resilience on the transactional queries it cannot complete, and no line in any report that says which is which. The response is triage: measure your own exposure query by query, protect the commercial pages, and repurpose the informational content whose click traffic is not coming back into the citation asset it can still be.

The traffic impact of AI Overviews

Google rolled AI Overviews out to all US users in May 2024 and to over a billion people worldwide by the end of that year, and the click data now has proper research behind it. Pew Research Center tracked the browsing of 900 US adults through March 2025 and found that when an AI summary appeared, users clicked a traditional result on 8 percent of visits, against 15 percent when it did not. Clicks on links inside the summary itself were close to non-existent, at 1 percent of visits.

The effect is not uniform. Queries where the overview satisfies the intent outright, simple factual questions and basic how-to content, see the sharpest declines. Queries where the buyer has to visit a site to finish the task, purchasing, comparing specific products, booking, hold up better because the overview cannot complete the transaction.

For most Shopify brands the commercial queries that matter fall into the second group. A buyer researching a specific type of boot or a premium baby brand is not completing the purchase inside an AI Overview, they are using it for category education and then clicking through to a brand or retailer they now trust.

Reading your own exposure

No report separates AI Overview traffic effects for you, since Search Console folds overview appearances into ordinary impressions, so the exposure read is manual and worth an afternoon.

1. Take your top fifty non-branded queries from Search Console and run each in a clean browser session, recording which trigger an overview and whether you or a competitor appears in it.
2. Sort the triggering queries by intent: educational queries the overview fully answers, comparison queries it partially answers, and buying queries it cannot complete.
3. In Search Console, watch the impressions-to-clicks relationship on the triggering queries: stable positions with falling click-through is the overview absorbing clicks above you, a pattern worth annotating so nobody misdiagnoses it as a rankings problem later.
4. Repeat quarterly with the same queries, because overview coverage by category is still expanding and the triage list moves.

The triage

The middle row is where the contest sits. A comparison overview that names three brands has effectively written the buyer's shortlist, and the next search is usually branded. Watching which brands your category's overviews name, and which sources they cite, tells you exactly whose content is framing your market.

Being cited in an AI Overview

Brands whose content is cited as a source in the overview keep a form of visibility even when users do not click, because the brand name appears inside the answer itself. For category-level queries that exposure has value beyond the click, and it shapes which names the buyer carries into the next, more commercial search.

The content most likely to be cited is content that answers the query with specific, verifiable information. Buying guides with detailed product comparisons, FAQ content with clear answers to specific questions, and structured data that helps Google connect the content to the query all raise the odds.

Google's own guidance on AI features points the same way, and is notably unexotic: there is no special markup for overview inclusion, and the advice reduces to standard indexability plus content worth citing. That makes overview citation an extension of the same evidence-dense content work that earns citations everywhere else, rather than a separate discipline with separate tricks.

What we see on live accounts

The Search Console pattern is consistent across the accounts we work on. Informational queries hold or grow impressions while click-through decays, often with positions unchanged, which is the overview's fingerprint: the page is still ranking, above an audience that now reads the answer instead. Meanwhile the transactional queries in the same accounts show ordinary volatility and nothing structural.

Two working practices follow. We annotate the date an overview first appears on a tracked query, because six months later somebody will ask what happened to that page's traffic and the answer should be on record rather than re-litigated. And we read branded search volume alongside the informational decline, since a brand being named in overviews and AI answers shows up as branded demand before it shows up anywhere else, and on several accounts the branded line has grown while the informational line fell. One buyer journey, measured at two different doors.

What this means for the traffic line

Report the shift plainly rather than absorbing it silently. Total organic clicks is now a blended metric that punishes exactly the informational content doing its new job well, so separate branded from non-branded, transactional from informational, and let each group carry its own target: revenue and rankings for the commercial set, citation presence and assisted branded demand for the educational set.

And resist the reflex conclusions. The traffic that AI Overviews absorb was the lowest-converting slice of organic, the buyers who click through past an overview arrive better educated, and the brands that panic-cut their informational content are removing themselves from the answers their buyers now read instead, so respond in proportion.