First, two numbers have been going around this year. AI traffic converts about 50% better than organic search. Yet AI-only shoppers convert worse than search-only shoppers. Both are true, which makes ai search roi harder to judge than either camp admits. The gap between them turns out to be the most useful thing a merchant can learn here.
More than half of AI-referred sessions land straight on a product page. Plug In SEO‘s Scanner checks whether those product titles and descriptions actually carry the keywords you are targeting.
The growth is real, and it is fast
Start with the part nobody disputes. So far the channel is genuinely expanding.
On Shopify storefronts, referral sessions from AI chatbots grew more than 8x year over year, and AI-referred orders grew nearly 13x. Across the wider web, AI appeared in 11.4% of shopper journeys in June 2026, up from 4.5% in July 2024. That growth shows up in AI Overviews on shopping searches too.
AI-referred visitors convert at nearly 50% higher rates than organic search visitors on product detail pages, and AI-attributed orders carry 14% higher average order values.
Those are strong numbers. But they need one more figure before they mean what they appear to mean.
Why AI traffic converts better than search traffic
Here is the figure that explains the rest. Still, it is in the same Shopify data and gets quoted far less.
More than half of AI-referred sessions start on a product detail page. But for organic search, that number is only about 20%.
So AI-referred shoppers are not a better class of customer. The assistant did the discovery work first, compared the options, and sent the shopper straight to the item. In other words, you are handed people at the end of their research, not the beginning.
That is a selection effect, not a channel effect. It matters, because it changes what you should do about it. The lever is not “get more AI traffic” — it is making sure a product page can survive being landed on cold, with no category page or homepage to set it up.
The counter-data, and why both numbers are true
Now the other measurement. However, it points the opposite way.
Similarweb and Statista tracked whole shopper journeys rather than last clicks. Measured that way, conversion rates run: AI-only 12.5%, Search-only 17.3%, AI plus Search 23.0%.
| Journey type | Conversion rate |
|---|---|
| AI only | 12.5% |
| Search only | 17.3% |
| AI + Search | 23.0% |
Read that carefully. Using AI on its own converts worse than using search on its own, and the combination beats both.
The two findings do not actually contradict each other, because they count different things. Shopify measures sessions arriving from an AI domain on the last click. Similarweb measures which tools appeared anywhere in the journey. A shopper who asks ChatGPT, then googles the brand, then buys is an organic-search session to one and an AI-plus-Search journey to the other.
Anyone quoting one of these numbers without the other is selling something. Keep them separate and both stay useful.
There is barely such a thing as an AI-only shopper
This is the finding that settles the budget question. It is also the least surprising one.
Of the journeys that involve AI, 89% also involve Search. Only 11% use AI without touching a search engine at all.
So a separate AI budget is mostly a budget for 11% of a channel that appears in 11.4% of journeys. Meanwhile the best-converting journey on the board is the one that uses both, and you cannot buy that one separately either.
What the honest ai search roi answer looks like
In short: worth attention. But not worth a separate line item.
The work that earns AI citations is overwhelmingly the same work that earns rankings. Substantive product pages, brand mentions on other people’s sites, clean structured data, pages that load. We went through what actually correlates with citations in how to get cited by ChatGPT, and almost none of it is AI-specific. The same holds for which sites assistants actually cite.
So the useful question is not how to split a budget. It is whether your product pages can carry a sale on their own, because more than half of your AI visitors will never see anything else. Our product page SEO guide covers what those pages need.
Three caveats worth stating plainly
But none of these kill the case. They just get left out of the confident versions.
- Sample sizes are not always published. Shopify describes its figures as Q1 2026 commerce data without disclosing how many merchants that covers.
- The direction reversed recently. AI-referred traffic converting 42% better than Google traffic is a genuine turnaround from a year earlier, when the same traffic converted 38% worse. A metric that flipped that hard in twelve months can flip again.
- Platform share estimates disagree wildly. Published figures for ChatGPT’s share of AI referral traffic range from roughly 57% to 88% depending on who measured and what counted. Treat any precise platform split with suspicion.
What to actually do this quarter
Short list, in order.
- First, open your three best-selling product pages and read them as a stranger with no context. That is the AI-referred experience.
- Fill in thin or missing product titles and descriptions before anything else, since those pages are the landing pages now.
- Keep doing normal SEO. It is the same work, and the combined journey converts best.
- Track AI referrals separately in analytics so you have your own numbers rather than the industry’s.
- Do not build a separate GEO budget on the strength of a metric that reversed twelve months ago.
Frequently asked questions
Is AI traffic worth chasing for a Shopify store yet?
It is worth attention, but not a separate budget. AI appeared in 11.4% of shopper journeys in June 2026, and 89% of those journeys also involved search, so there is barely an AI-only audience to target. The work that earns AI citations is largely the same work that earns rankings.
Why does AI traffic convert better than organic search?
Mostly because of where it lands. More than half of AI-referred sessions start on a product detail page, against about 20% for organic search, so the assistant has already done the discovery step. The shopper arrives at the end of their research rather than the start, which is a selection effect rather than a quality difference.
Do AI-only shoppers convert better than search shoppers?
No. Measured across whole journeys, AI-only converts at 12.5% against 17.3% for search-only, while journeys using both convert best at 23.0%. That sits alongside the last-click figures showing AI traffic converting higher, because the two studies count different things.
How much should I spend on AI search optimisation?
Treat it as part of your SEO budget rather than a new one. The highest-value actions — substantive product pages, off-site brand mentions, clean structured data — serve both channels at once. Spend the effort on product pages specifically, since those are where AI-referred shoppers actually land.
Which AI platform sends the most ecommerce traffic?
ChatGPT leads by a wide margin, but be careful with precise figures. Published estimates of its share of AI referral traffic range from roughly 57% to 88% depending on the measurement, and that spread is a good reason to track your own analytics rather than relying on industry averages.
Could these numbers change?
They already have. AI-referred traffic now converts around 42% better than Google traffic, reversing a position a year earlier when it converted 38% worse. A metric that swung that far in twelve months deserves rechecking before you build a strategy on it.
Your product pages are the landing pages now
Plug In SEO’s Scanner checks whether your product titles and descriptions carry the keywords you are targeting, so the pages AI sends shoppers to are working for the right searches.