Vertical Market Evolution
The Agentic Shopper Premium: Measuring the AI-Order Effect in Ecommerce
Shoppers who arrive at your store from an AI answer convert better and spend more than shoppers who arrive from a search result, across every measurement stack that has looked, but the channel is still small and every number so far comes from a vendor grading its own traffic.
Part of Vertical Playbooks in the Insights library.
Abstract
Four independent analytics stacks, Shopify, Adobe, Salesforce, and Similarweb, all report the same shape of finding: shoppers referred by an AI engine convert at meaningfully higher rates, spend more per order, and skip straight to product pages far more often than shoppers referred by organic search. That pattern is well evidenced. What is not yet evidenced is why: whether it reflects genuinely higher purchase intent or is partly an artifact of who is using AI shopping tools today and how they land on a page, and whether "agentic" is even the right word for what is mostly still AI-assisted research followed by a human at checkout.
AI-referred traffic to ecommerce sites is one of the fastest-moving lines in retail analytics right now, and the growth curve is not subtle. Adobe, tracking over a trillion visits to US retail sites, measured AI-referred traffic up 393% year over year in the first quarter of 2026, and by June 2026 was reporting a cumulative gain of 1,324%, more than fourteen times, since it started tracking the channel in October 2024. Shopify's own merchant network saw AI-referred orders grow almost 13 times year over year in the same window. These are steep curves off what was, until recently, a channel that barely registered. The more interesting shift is qualitative rather than just volumetric: Adobe found that in March 2025, AI-referred traffic converted 38% worse than everything else on a retail site. One year later, in March 2026, that same channel converted 42% better. That is not a channel slowly warming up. That is a full reversal in the space of twelve months, and it is the clearest sign in the evidence that AI referral went from a novelty click to a meaningful part of how people shop in a single year.
The behavioral signature is consistent enough across vendors that it is worth taking seriously even before the causal question is settled. Shopify found AI-referred sessions convert nearly 50% higher than organic search on product pages, beating organic in 23 of 25 merchant categories by an average of 56%, and that more than half of AI-referred sessions land directly on a product detail page versus about 20% for organic search. Similarweb, looking at a different panel entirely, found ChatGPT-referred traffic converting at 11.4% against 5.3% for organic search, better than double, and described the pattern as shoppers arriving primed to buy because they had already done their comparison shopping inside the AI tool before they ever hit your site. Adobe's own commissioned survey of over 5,000 US shoppers backs this up from the demand side: 39% said they already use AI as part of how they shop, 85% said it improved the experience, and 66% said they trust the AI tool's results. Put together, the picture is of a shopper who has done more of the decision-making work before arriving, which is exactly the kind of visitor every retailer wants more of. What none of this yet proves is whether that is because AI referral genuinely creates more intent, or because the AI-referred audience today skews toward people who were always going to convert well.
This is a textbook case for reading the terrain rather than trusting a single vendor's number. Every headline figure in this study (Shopify's 13x, Adobe's 393% and 138%, Salesforce's 20% of holiday sales) is proprietary analytics measured by that company on its own merchant or client base, not an independent, market-wide audit. And the dramatic percentages are being computed off a base that independent measurement puts in the low single digits of total traffic: an analysis of 200 Stripe-connected merchant sites found AI engines driving a median of roughly 6% of sessions as of May 2026, and standard GA4 analytics is misreading a large share of that as ordinary Direct traffic rather than AI referral, which means many stores are already receiving this premium and don't know it. The read is that AI referral is a real and rapidly growing part of where retail attention now lands, but it is not yet the majority of the terrain, and no one, including the platforms reporting these numbers, has independently reconciled what percentage of a given business's actual demand now runs through this surface versus classic search or direct visits. That is precisely the gap a Visibility Corpus for your category is built to close: where your buyers' attention actually sits today, measured against your own site and your own competitors.
The data, in one read
The numbers everyone is now citing
Four analytics organizations, working from four different datasets, have independently landed on the same basic finding this year: shoppers who arrive at a retail site from an AI engine behave differently, and better, than shoppers who arrive from a traditional search result. Shopify, working from its own merchant network, reported AI-referred orders up nearly 13 times year over year as of May 2026, with those orders carrying a 14% higher average order value than orders from organic search. Adobe, working from an analysis of more than a trillion visits to US retail sites, measured AI-referred traffic up 393% year over year in the first quarter of 2026 alone, and by June 2026 put the cumulative growth at 1,324% since it started tracking the channel in October 2024. Salesforce, working from its holiday shopping index, said AI influenced roughly 20% of global online retail sales during the 2025 holiday season, an estimated $262 billion in commerce. None of these four organizations shares a dataset, a definition, or a methodology with the others, and yet they all point the same direction: this is a real, fast-growing part of how people shop now, not a rounding error or a press-release artifact.
Four analytics organizations, working from four different datasets, have landed on the same finding this year: AI-referred shoppers behave differently, and better.
What an AI-referred shopper actually does differently
The growth numbers get the headlines, but the behavioral numbers are more useful for a business owner trying to decide what to do about this. Shopify found AI-referred sessions converting nearly 50% higher than organic search on product pages, and outperforming organic search in 23 of 25 merchant categories by an average of 56%. More tellingly, Shopify found that more than half of AI-referred sessions, roughly 55% by its own count, land directly on a product detail page, compared with about 20% for organic search traffic. That is a meaningfully different shopper journey: instead of arriving on a homepage or a category page and browsing toward a decision, the AI-referred visitor has often already decided roughly what they want and is arriving to confirm it and buy. Similarweb, working from an entirely separate traffic panel, found the same shape of result: ChatGPT-referred visitors converting at 11.4%, more than double the 5.3% rate for organic search, and described those visitors as arriving already primed to buy because they had done their research inside the AI tool before ever landing on the site.
More than half of AI-referred sessions land straight on a product page. About one in five organic-search sessions do the same.
A channel that flipped in twelve months
The single most striking number in the evidence is not a growth rate, it is a reversal. Adobe found that in March 2025, AI-referred traffic converted 38% worse than the rest of a retail site's traffic, consistent with a channel still full of curious clickers and comparison browsers who weren't ready to buy. One year later, in March 2026, that same channel converted 42% better than everything else, alongside 37% higher revenue per visit, 48% more time spent on-site, and 13% more pages viewed per visit. A channel does not usually move from meaningfully worse than average to meaningfully better than average in a single year unless something structural has changed in who is using it and why. That doesn't prove the premium is permanent (channels have compressed toward the mean before as they go mainstream), but it is a strong signal that the early skepticism about AI-referred traffic being low-quality curiosity clicks no longer matches what the traffic is doing.
What holiday 2025 showed at scale
Salesforce's holiday index is the largest single dataset in this evidence base, and it adds two findings worth sitting with. First, shoppers referred from third-party AI search channels such as ChatGPT and Perplexity converted 9 times more often than shoppers referred from social media, a gap large enough to suggest AI referral and social referral are simply not comparable channels for purchase intent. Second, retailers who had deployed their own AI shopping or service agents (Salesforce names Pandora, SharkNinja, and Funko) grew holiday sales 59% faster than retailers who hadn't: 6.2% year-over-year growth versus 3.9% for non-adopters. Salesforce's own language matters here: the company describes this volume as influenced by AI, not autonomously transacted by AI and agents. That distinction matters and we hold to it throughout this study, because conflating the two is exactly the kind of overstatement that erodes trust once a business owner checks the underlying claim against its own analytics.
The percentages are real, the base is still small
Here is the part of the story vendor press coverage tends to skip. An analysis of 200 Stripe-connected merchant sites, run by a company called Attrifast that sells a product built to detect exactly this kind of traffic, found AI engines driving a median of roughly 6% of total sessions as of May 2026, a figure we tier as contested precisely because the publisher has a direct commercial interest in the number looking large. The same analysis found that standard GA4 analytics undercounts true AI-referral traffic by a median of 64%, largely because AI referrals get misclassified as ordinary Direct traffic, which means a meaningful share of stores are almost certainly already receiving AI-referred visitors and have no clean way to see it in their own dashboard. eMarketer, in a forecast we tier as emerging because the underlying figures sit behind a paywall we could not fully re-verify line by line, projects US ecommerce sales transacted directly through AI platforms will exceed $20 billion in 2026, about 1.5% of US ecommerce, rising to $144 billion, or 8.8%, by 2029. Reconciling that 1.5%-transacted figure with Salesforce's 20%-influenced figure requires knowing exactly how each firm defines an AI sale, and the evidence does not yet let us do that cleanly. Those are almost certainly measuring different things (influence versus direct transaction), but the size of the gap between them is itself a reminder that no two of these headline numbers should be read as directly comparable.
A quarter of your traffic report may already be AI-referred and simply logged as Direct.
"Agentic" oversells what is actually happening today
This study's title borrows the word agentic because that is the word the industry is using, but the evidence pushes back on it. Forrester's mid-2026 analysis of agentic commerce found that most of what gets called agentic in practice is still conversational: chat-based product discovery and recommendation is widespread, but a human is still driving the purchase decision and completing checkout in most cases, and consumers rarely let an AI agent finish a purchase without direct oversight. In other words, the growth we are measuring across Shopify, Adobe, Salesforce, and Similarweb is overwhelmingly an AI-assisted research effect that ends in a person clicking buy, not an autonomous agent transacting on a shopper's behalf. Bain and Company's forward-looking estimate, that the US agentic commerce market (purchases initiated, influenced, or completed by AI agents) could reach $300 to $500 billion by 2030, roughly 15% to 25% of US ecommerce, is a genuinely useful number, but it is explicitly a forecast from a single firm, not a present-day measurement, and we tier it emerging accordingly. The right framing for a business owner today is: build for AI-assisted research that ends in a human checkout, not for autonomous agents placing orders on your customers' behalf. That second thing is coming, based on Bain's estimate, but it is not what is happening at scale yet.
What we still don't know
The gap that matters most is causal. Nothing in this evidence isolates whether the AI-referred conversion and order-value premium comes from genuinely higher purchase intent, or from confounds that would inflate the numbers regardless of intent: shoppers who land directly on a product page convert better than shoppers who land on a homepage no matter how they got there, and the population currently using AI shopping tools likely skews toward higher-income, more AI-literate, early-adopter buyers who would convert well through any channel. No controlled study or econometric analysis in the public record separates those effects out. There is also no independent, cross-platform audit of AI-referred order volume; every headline figure in this study (Shopify's, Adobe's, Salesforce's) comes from that company's own client base, directionally consistent with the others but not strictly auditable against a shared standard. And nearly every number covers a single year of tracking, so it is not yet possible to say whether this premium is a durable feature of AI-referred demand or an early-adopter effect that will compress toward parity as the channel becomes mainstream, a pattern seen historically with other new high-intent channels as they scale.
What this means for your business right now
None of this is a reason to rebuild your store around autonomous agents. The evidence doesn't support that as today's reality. It is a reason to check three things. First, look at how much of your own traffic is currently being logged as Direct, given that GA4 is known to misclassify a large share of genuine AI referral that way; you may already be receiving this premium and not know it. Second, make sure your product pages can stand alone as an entry point, since more than half of AI-referred visitors across the evidence arrive there directly rather than working through your homepage. Third, treat every headline percentage in this study as directionally true and numerically vendor-specific: track your own AI-referred conversion rate over time rather than assume an average computed on someone else's merchant base applies to your store. This channel is growing fast off a small base, and the businesses that benefit most will be the ones measuring their own version of it early, not the ones reacting to a trend piece a year from now.
The evidence, in numbers
Key findings, dated and sourced
-
AI-referred orders on Shopify grew nearly 13x year over year in Q1 2026, and orders attributed to AI-powered search carried 14% higher average order value than orders from organic search.
established AI-referred shoppers convert better and spend more, Shopify Enterprise Blog (2026-05-11)
-
AI-referred sessions on Shopify storefronts convert nearly 50% higher than organic search on product detail pages, outperforming organic search in 23 of 25 merchant categories by an average of 56%. More than half of AI-referred sessions land directly on a product detail page, versus about 20% for organic search.
established AI-referred shoppers convert better and spend more, Shopify Enterprise Blog (2026-05-11)
-
AI-referred traffic to US retail websites rose 393% year over year in Q1 2026, based on analysis of over a trillion visits to US retail sites, with the trailing twelve months to March 2026 up 269% year over year.
established AI traffic to US retailers rose 393% in Q1, and it's boosting their revenue too, TechCrunch, reporting Adobe Analytics data (2026-04-16)
-
In March 2026, AI-referred traffic converted 42% better than non-AI traffic on US retail sites, a full reversal from March 2025 when the same channel converted 38% worse than non-AI traffic. AI-referred visitors also generated 37% higher revenue per visit, spent 48% longer on-site, and viewed 13% more pages per visit than non-AI traffic.
established AI traffic to US retailers rose 393% in Q1, and it's boosting their revenue too, TechCrunch, reporting Adobe Analytics data (2026-04-16)
-
AI-referred traffic to retail sites grew 138% year over year in May 2026, and has grown a cumulative 1,324%, more than 14x, since Adobe began tracking the channel in October 2024.
established Adobe: AI-referred traffic to retail sites doubles in a year, Digital Commerce 360, reporting Adobe Analytics data (2026-06-17)
-
During the 2025 holiday shopping season, Salesforce reports AI influenced roughly 20% of global online retail sales, an estimated $262 billion in commerce. Shoppers referred from third-party AI search channels such as ChatGPT and Perplexity converted 9 times more often than shoppers referred from social media. Salesforce's own language is that AI influenced this volume, not that AI or agents autonomously transacted it.
established Salesforce Reveals 2025 Holiday Shopping Data, Salesforce Shopping Index (2026-01)
-
Retailers that deployed their own AI shopping or service agents, including Pandora, SharkNinja, and Funko, grew holiday 2025 sales 59% faster than retailers that had not, averaging 6.2% year-over-year growth versus 3.9% for non-adopters.
established Salesforce Reveals 2025 Holiday Shopping Data, Salesforce Shopping Index (2026-01)
-
eMarketer forecasts US ecommerce sales transacted directly through AI platforms, including ChatGPT, Gemini Shopping, and Perplexity, will exceed $20 billion in 2026, about 1.5% of US ecommerce, and top $144 billion, about 8.8%, by 2029. The underlying figures sit behind a paywall and could not be independently re-confirmed line by line beyond the published chart headline.
emerging US Ecommerce Sales via AI Platforms Will Exceed $20 Billion in 2026, Top $144 Billion by 2029, eMarketer chart
-
Similarweb's Global Ecommerce Report found ChatGPT-referred site traffic converts at 11.4%, versus 5.3% for organic search, a 2.15x difference, framing AI-referred visitors as arriving already primed to buy after doing their research inside the AI tool.
established Similarweb's 3rd Annual Global Ecommerce Report: Growth Shifts to Apps and AI (2025-09-15)
-
An analysis of 200 Stripe-connected merchant sites found AI engines drove a median of about 6% of total sessions as of May 2026, and that standard GA4 analytics undercounts true AI-referral traffic by a median of 64% because those referrals are frequently misclassified as Direct traffic. The publisher sells a paid product built to detect this exact measurement gap, a direct conflict of interest that discounts but does not disqualify the figure.
contested Attrifast, Is AI Traffic Actually Worth It? An Honest 2026 Answer (with Data) (2026-05)
-
Most agentic commerce experiences in mid-2026 remain conversational rather than autonomous: chat-based product discovery and recommendation is widespread, but humans still drive purchase decisions and checkout in most cases, and consumers rarely let an AI agent complete a purchase without direct oversight.
established The State Of Agentic Commerce In Mid-2026, Forrester Blog (2026-05-28)
-
Bain and Company estimates the US agentic commerce market, purchases initiated, influenced, or completed by third-party or retailer-hosted AI agents, could reach $300 to $500 billion by 2030, roughly 15% to 25% of total US ecommerce. This is a forward-looking market-sizing estimate from a single firm, not a current measurement.
emerging 2030 Forecast: How Agentic AI Will Reshape US Retail, Bain & Company Snap Chart (2025-12-17)
-
In an Adobe-commissioned survey of over 5,000 US respondents, 39% said they used AI for online shopping, 85% said it improved their shopping experience, and 66% said they believe AI tools provide accurate shopping results.
established AI traffic to US retailers rose 393% in Q1, and it's boosting their revenue too, TechCrunch, reporting Adobe survey data (2026-04-16)
Learning outcomes
What this study teaches
- AI-referred shoppers convert better and spend more across every measurement stack that has looked at it, including one comparison that flipped from 38% worse than average to 42% better than average in a single year, but this is still measured by vendors on their own traffic, not by an independent auditor.
- Standard analytics tools likely undercount this traffic today. One analysis found GA4 misclassifying a large share of genuine AI referrals as ordinary Direct traffic, meaning you may already be receiving this premium without seeing it in your dashboard.
- "Agentic" oversells what is happening in mid-2026. Forrester's own research finds most of this volume is AI-assisted research that still ends with a human completing checkout, not an autonomous agent placing the order, so build for that handoff, not for agent-run checkout.
- More than half of AI-referred visitors across the evidence arrive directly on a product page rather than a homepage, so a product page that can stand alone, answer the question, and close the sale matters more for this channel than it does for browsers arriving through a homepage.
- Every headline number in this space, including the ones in this study, comes from a vendor's own client base and covers roughly a year of tracking. Treat the direction as credible and the magnitude as vendor-specific, and measure your own store's AI-referred performance over time rather than assume an industry average applies to you.
Honest limits
What this does not yet settle
- No independent, cross-platform audit or market-wide dataset yet measures AI-referred orders or agentic commerce volume. Every headline growth figure in this study (Shopify's roughly 13x, Adobe's 393% and 138%, Salesforce's 20% of holiday sales) is proprietary analytics from that company's own merchant or client base, directionally consistent with the others but not strictly comparable or independently auditable.
- No causal research yet isolates whether the AI-referred conversion and order-value premium reflects genuinely higher purchase intent, or confounds such as disproportionate direct-to-product-page entry, which tends to convert better through any channel, and a demographic skew toward higher-income, more AI-literate early adopters among current AI shoppers.
- The large growth percentages (13x, 393%, 138%, 1,324%) are computed off a base that independent measurement suggests is still in the low single digits of total ecommerce sessions, a piece of context that is largely absent from how these numbers get reported.
- Salesforce's figure of AI influencing about 20% of global online sales and eMarketer's figure of AI platforms directly transacting about 1.5% of US ecommerce cover roughly the same period but are difficult to reconcile without knowing precisely how each firm defines an AI sale. This is a genuine, unresolved definitional gap, not a rounding difference.
- No public research of the depth normally produced for on-site checkout and cart behavior was found comparing how AI-chat-referred shoppers actually navigate a store's funnel against organic-search shoppers. Shopify's product-page-entry statistic is the closest available proxy, but it stands uncorroborated by any independent source.
- Nearly every figure in this study covers a single year of tracking, so it is not yet possible to say whether the conversion and order-value premium is a durable feature of AI-referred demand or an early-adopter effect that will compress toward parity as the channel becomes mainstream, the pattern seen historically with other new high-intent channels as they scale.
This is a synthesis of dated, attributed evidence, not a census. The AI-answer layer in particular has no independent, Nielsen-grade measurement yet, so readings of it are directional and named as a frontier, never presented as settled.
Straight answers
Frequently asked questions
Do shoppers referred by AI engines actually convert better than shoppers from regular search?
Yes, across every measurement stack that has looked. Shopify found AI-referred sessions converting nearly 50 percent higher than organic search on product pages, and Similarweb found ChatGPT-referred traffic converting at 11.4 percent versus 5.3 percent for organic search, more than double. Adobe measured a similar swing at the whole-channel level: AI-referred traffic on US retail sites converted 42 percent better than non-AI traffic in March 2026. None of these figures come from an independent auditor, they are each a vendor measuring its own traffic, but four separate companies landing on the same direction is a meaningful signal.
Could some of our AI-referred traffic already be showing up as 'Direct' in our analytics?
It is plausible. An analysis of 200 Stripe-connected merchant sites found that standard GA4 analytics undercounts true AI-referral traffic by a median of 64 percent, largely because those referrals get misclassified as ordinary Direct traffic. That figure is tiered contested in this study because the publisher, Attrifast, sells a product built to detect exactly this gap, which is a direct conflict of interest that discounts but does not disqualify the finding.
Does 'agentic commerce' mean AI agents are now placing orders without a human?
Not yet, at least not at scale. Forrester's mid-2026 research on agentic commerce found that most of what gets called agentic in practice is still conversational: a human is driving the purchase decision and completing checkout in most cases, and consumers rarely let an AI agent finish a purchase without direct oversight. The growth this study measures is overwhelmingly an AI-assisted research effect that ends with a person clicking buy, not autonomous agents transacting.
How big is AI-referred traffic as a share of total ecommerce traffic right now?
Still small, even though the growth rates look dramatic. The same Attrifast analysis of 200 Stripe-connected merchant sites found AI engines driving a median of roughly 6 percent of total sessions as of May 2026. Separately, eMarketer, in a figure this study tiers emerging because the underlying data sits behind a paywall, projects AI platforms will directly transact about 1.5 percent of US ecommerce in 2026, rising to 8.8 percent by 2029. The large percentage gains reported by Shopify, Adobe, and Salesforce are all computed off that still-small base.
Will the AI-referred conversion premium hold up, or is it likely to fade as the channel matures?
That is not yet known. Nearly every figure in this study covers a single year of tracking, so there is no way yet to tell whether the premium is a durable feature of AI-referred demand or an early-adopter effect that compresses toward parity as the channel becomes mainstream, a pattern seen historically with other new high-intent channels as they scaled. No causal research has isolated whether the premium reflects genuinely higher purchase intent or is partly explained by who currently uses AI shopping tools and how they land on a page.
Provenance
References
- Shopify Enterprise Blog, "AI-referred shoppers convert better and spend more," May 11, 2026 https://www.shopify.com/enterprise/blog/ai-search-insights
- TechCrunch, reporting Adobe Analytics data, "AI traffic to US retailers rose 393% in Q1, and it's boosting their revenue too," April 16, 2026 https://techcrunch.com/2026/04/16/ai-traffic-to-us-retailers-rose-393-in-q1-and-its-boosting-their-revenue-too/
- Digital Commerce 360, reporting Adobe Analytics data, "Adobe: AI-referred traffic to retail sites doubles in a year," June 17, 2026 https://www.digitalcommerce360.com/2026/06/17/adobe-ai-referred-traffic-to-retail-sites-doubles-in-a-year/
- Salesforce, "Salesforce Reveals 2025 Holiday Shopping Data," Salesforce Shopping Index, January 2026 https://www.salesforce.com/news/stories/2025-holiday-shopping-data/
- eMarketer, "US Ecommerce Sales via AI Platforms Will Exceed $20 Billion in 2026, Top $144 Billion by 2029," 2026 https://www.emarketer.com/chart/c/358389/us-ecommerce-sales-via-ai-platforms-will-exceed-20-billion-2026-top-144-billion-by-2029-358389
- Similarweb, "3rd Annual Global Ecommerce Report: Growth Shifts to Apps and AI," September 15, 2025 https://ir.similarweb.com/news-events/press-releases/detail/132/similarwebs-3rd-annual-global-ecommerce-report-growth-shifts-to-apps-and-ai
- Attrifast, "Is AI Traffic Actually Worth It? An Honest 2026 Answer (with Data)," May 2026 https://attrifast.com/blog/is-ai-traffic-worth-it
- Forrester Blog, Emily Pfeiffer, "The State Of Agentic Commerce In Mid-2026," May 28, 2026 https://www.forrester.com/blogs/the-state-of-agentic-commerce-in-mid-2026/
- Bain & Company, Aaron Cheris, Mikey Vu, Stephanie Koszyk, Katherine Hall, "2030 Forecast: How Agentic AI Will Reshape US Retail," Snap Chart, December 17, 2025 https://www.bain.com/insights/2030-forecast-how-agentic-ai-will-reshape-us-retail-snap-chart/
Every measured figure is dated to its capture and tagged with an evidence tier. Every cited work is real and locatable. Where an engine could not be captured this round, it is named as uncaptured, not estimated. Small-sample readings are labelled as directional.