Amazon Prime Day ran June 23–26. The consumer basket numbers were mixed — average order size fell to $47.66, down from $53.34 last year, and household spend dropped roughly $13. But Adobe tracked something else in the traffic data: AI-referred visits to US retail sites grew 98.3% year over year on Day 1 alone.
Those shoppers converted 50.7% better than visitors from non-AI channels. They spent 49.9% more time on product pages, viewed 20.5% more pages, and added to cart 33% more often. Adobe had projected 103% growth in AI-driven retail clicks across the full four-day event.
Why AI shoppers behave differently
When someone opens Perplexity and types "best trail running shoes under $120 for wide feet" and clicks through, they have already done the consideration work. The AI tool filtered, ranked, and recommended before the click happened. By the time they land on a product page, they're checking one option — not browsing ten.
That's the mechanism behind the 50.7% conversion premium. These aren't awareness shoppers. They arrive later in the funnel, with higher specificity, and they're on your page to confirm what an AI system already told them to expect.
The add-to-cart rate follows the same logic: 33% higher because the intent driving the session is purchase intent, not curiosity.
This pattern has been building. AI traffic to US retailers grew 393% in Q1 2026, per Adobe data published in April. Prime Day brought it into sharp relief because the volume and the stakes were both higher.
The readability gap most brands didn't close
Adobe also flagged something uncomfortable in the same report: large portions of retail websites remain unreadable by machines.
Perplexity, ChatGPT Shopping, Google AI Mode, and Claude pull product information from structured data and crawlable HTML. If your PDPs rely on JavaScript-rendered specs, bury key attributes in images, or lack Product schema — name, price, availability, description, image — AI tools either skip your products or surface an incomplete version. The shopper never sees you.
The practical effect during Prime Day: brands with clean, machine-readable product data captured the better buyer. Everyone else got the traditional traffic — which, per Numerator, came with smaller orders.
That's the real split in the Prime Day data. The headline consumer numbers looked soft. But buried inside them is a segmentation story: some brands captured a high-intent, AI-routed buyer who converted at a meaningfully higher rate. Others competed for the same shrinking average-order-value shopper they've always had.
What to fix before Q4
Q4 planning starts in earnest this month. The stretch between now and October is when you have time to fix machine readability without doing it under deadline pressure.
Start with visibility. Search for your top five SKUs in Perplexity and ChatGPT with buying-intent queries — "best [product] under $X for [use case]," "which [product] works with [compatible item]." If your products don't appear, or appear with wrong pricing and missing attributes, you have a gap.
Product schema is the first lever. Every PDP needs structured data with name, price, offers.availability, description, and image. Without it, AI crawlers can't reliably extract what you sell. Text-based specs come next — dimensions, materials, compatibility details, and differentiators should be in readable paragraph or list copy, not locked inside product images or rendered dynamically via JavaScript. AI systems read HTML, not photography.
Feed freshness is the third variable. If you run Google Shopping or Meta catalogs, your feed accuracy determines what AI Mode and Advantage+ Shopping surface to shoppers. Stale descriptions, missing attributes, and out-of-stock items in active feeds degrade both paid performance and AI discoverability at the same time.
None of this is novel SEO work. It's product data hygiene that the shift to AI-mediated discovery has made compulsory rather than optional.
If you want a clear read on where your product pages stand on the machine-readability spectrum right now, the free audit at Gromerce surfaces the gaps in a few minutes.
Prime Day ended two days ago. The brands that captured AI traffic got better buyers. The ones that didn't have a quarter to fix it before the same pattern plays out at higher Q4 CPMs.
Sources: Adobe Digital Insights, Numerator Prime Day 2026 Tracker, Retail Brew, Chain Store Age, PMG Prime Day 2026 Day 3 Recap, TechCrunch, June 2026
