AI-Referred Shoppers Are Now Your Best Customers

"Companies have spent decades refining consumer journeys, fine-tuning every click, scroll, and tap. But in the era of agentic commerce, the consumer no longer travels alone. Their digital proxies now navigate the commerce ecosystem, making millions of microdecisions daily. To thrive, brands must rethink the full stack of engagement," Coggins writes, "not for the people they've worked to understand but for the agents now acting on their behalf."

That observation came from Becca Coggins, a senior partner at McKinsey. She published it in "The agentic commerce opportunity," a McKinsey QuantumBlack piece from October 17, 2025. At the time, traffic arriving via AI shopping assistants was marginal for most brands: real and growing, but not yet large enough to reshape how anyone thought about acquisition priorities.

That assessment has become outdated.

The numbers changed direction

Adobe Digital Insights tracked a 393% year-over-year surge in traffic arriving via AI assistants through Q1 2026. Volume at that scale would command attention by itself. The more consequential finding sits alongside it: these shoppers converted at a rate 42% higher than the baseline (Adobe Digital Insights, April 2026).

Shopify's own first-party data adds texture to the Adobe picture. Risley, writing in the Shopify Enterprise Blog in May 2026, found that agent-driven buyers converted 49% above organic and arrived with a 14% premium on average order value. The category analysis reinforced the pattern: this cohort outperformed in 23 of 25 product categories tested. All five figures together:

Metric	Figure	Source
YoY traffic growth via AI assistants	+393%	Adobe Digital Insights, Q1 2026
Conversion lift vs. site baseline	+42%	Adobe Digital Insights, Apr 2026
Conversion lift vs. organic	+49%	(a)
Average order value premium	+14%	(a)
Category wins vs. non-agent buyers	23 of 25	(a)

(a) Risley, Shopify Enterprise Blog, May 2026. The two percentages use different denominators: Adobe's +42% compares to the overall site rate, while Shopify/Risley's +49% compares to organic visits specifically.

Traffic volume and purchase quality rising together is unusual. Channel growth typically introduces noisier intent: more people, including more who were never going to buy. The Q1 data runs counter to that pattern in both dimensions at once, which makes the combination worth understanding structurally rather than treating as a pleasant coincidence.

Why the channel inverted

The explanation lives in how AI assistants shop.

When a person delegates a purchase to an assistant, the assistant does the research first. It reads descriptions, parses specifications, and weighs alternatives against the buyer's stated preferences, filtering everything against the user's constraints before surfacing a recommendation. By the time a shopper lands on a product page via that path, the deliberation has finished. The work moved earlier in the process.

That is a structurally different kind of referral. Organic search sends visitors across the full spectrum of purchase intent: early-stage explorers, comparison shoppers, and buyers who are ready to transact all land in the same channel, and the aggregate conversion rate reflects that wide spread of motivations. An assistant referral compresses the population. The pre-purchase work happens before the session even begins, which makes the resulting traffic unusually homogeneous in intent. A ranked result refers a prospect. An assistant recommendation refers a buyer.

Journey compression: the PDP becomes the front door

Alongside conversion and AOV data, the Shopify Enterprise Blog reports a finding with direct architectural consequences: shoppers arriving via AI assistants are 2.75 times more likely to land on a product detail page than buyers coming from other channels.

Traditional ecommerce funnels assume an entry point through the homepage or a category view, narrowing through filters toward the PDP. Conversion optimization has historically treated that detail page as a late-stage asset, something to tune after awareness and consideration are solved.

Journey compression overturns that sequence. When a far larger share of your highest-converting traffic lands directly on the product page, the PDP stops functioning as the funnel's terminus.

It becomes the entry point.

That shift carries real implications for where merchandising attention goes. The homepage, the category experience, the editorial layer that supports awareness: all of that still matters for the channels that travel the traditional path. For assistant-driven buyers, though, the product page must carry different weight. It needs to orient a first-time visitor, confirm the recommendation that brought them there, and supply enough to close the transaction without assuming any prior browsing context.

A PDP designed for someone who already filtered by color, size, and price across three category pages handles a fundamentally different task than one receiving a cold landing. Most brands built their pages for the first scenario. Their best-converting buyers increasingly arrive through the second.

The practical redesign question runs deeper than layout. A person arriving via an AI assistant may have queried something like "a lightweight winter jacket that works in rain and for layering." The AI matched your product to that query and sent them to the PDP. If the page confirms the waterproofing rating, the fill weight, and how it layers over a midlayer, the purchase completes. If the copy says "versatile style for all occasions," the shopper has to re-engage the AI to verify, or they leave. Journey compression accelerates both outcomes: a well-matched page converts faster, and an opaque one loses the sale earlier than the traditional funnel would have.

What the inversion means for investment priorities

The reallocation case follows directly from the data. A channel delivering the best conversion rates, a 14% order value premium, and wins across 23 of 25 categories while growing 393% year over year deserves first-class investment. Agent optimization currently sits below SEO, paid search, and email in most brands' priority stacks, partly from habit and partly because the measurement ecosystem lags the volume data.

The straightforward move is to invest where the best buyers come from. Treating the assistant channel as secondary means systematically directing resources away from the acquisition source with the highest return. That mismatch is partly invisible because the channel is still small enough in absolute terms to feel optional. The 393% growth rate changes that calculation every month.

The implications cut across functions. Merchandising needs to rethink the PDP as a first-touch experience. Content functions need to map what information assistants use when forming recommendations, and identify where gaps in product data weaken those recommendations. Commerce architecture needs to assess whether the current stack surfaces the signals assistants read and keeps them accurate at scale.

None of that is optimization for a hypothetical future state. The inversion is already in the Q1 2026 data. Brands adjusting priorities now are responding to what has already happened.

The catch the data surfaces

Better conversion rates from assistant-driven traffic carry an embedded assumption: the assistant successfully matched the right product to the buyer. The 42% and 49% figures measure buyers who arrived. Nothing in the data accounts for buyers whose assistants could not form a confident recommendation and never sent them anywhere at all.

That gap points to a structural problem most brands have not yet addressed. AI assistants build recommendations from the product information they can access and parse. Discovery protocols like UCP (the Universal Commerce Protocol, built by Google and Shopify with retailers) make catalog items visible at scale, which is a prerequisite. But visibility alone does not determine whether a given item earns a recommendation over a competitor's offering.

Earning the recommendation depends on the assistant's ability to match a specific product to a buyer's stated intent. That matching requires authored reasoning: why this jacket fits the described build, how this item compares to the near-substitute the buyer mentioned, what the return policy means for a gift purchase, which size variant resolves the query. Platform product feeds and discovery protocols do not carry this reasoning, as a field audit of 2,483 Shopify merchants confirmed across every brand in the set. It lives in the merchant's head, in the brand's institutional understanding of the catalog, and in the buying guides most storefronts write for human readers rather than for algorithmic shoppers.

Brands that structure this reasoning at the product level give assistants real signal to work with. Those that rely on a discovery protocol alone give assistants visibility into a product but not a reason to prefer it. The inversion in conversion rates reflects the quality of today's referrals. The open question is how many referrals are not being made because a product is visible but opaque to the assistant deciding what to recommend.

Journey compression lands those high-intent buyers directly on the detail page, where the listing copy becomes the agent's primary reasoning substrate for deciding whether the item earns its place in a recommendation. And the richer that authored reasoning is, covering fit, substitution logic, return policy, and voice, the more likely an assistant surfaces the full brand-truth layer that separates a confident recommendation from a vague one.

What to do with this

The data from Adobe and Shopify points toward a concrete reordering of priorities. Invest in the assistant channel as a primary acquisition surface. Redesign the PDP as a first-impression experience. Close the gaps in authored reasoning that prevent your products from earning confident recommendations.

If you want to understand where your store sits on agent-readiness, including where the authored reasoning gaps are largest, we evaluate stores and keep an early access list for brands ready to work on this directly.

The buyers in the Q1 data who arrived via AI assistants were the best-performing customers those brands received that quarter. The worthwhile question is whether your store is prepared to receive them.

---

Sumit Jagdale is a co-founder of Sartorial. Sartorial builds the brand truth layer that makes merchant reasoning legible to AI shopping agents.