AI shoppers convert better, when the product page holds up

A year ago, a visitor sent by an AI assistant was one of the worst-converting visitors a store could get. Now they're one of the best. The change is worth understanding, because it shows where the work on AI search pays off.

The reversal

Adobe tracks more than a trillion visits to US retail sites. In May 2026, traffic from AI sources converted 54% better than traffic from everywhere else. A year earlier it converted at nearly half the rate. AI visitors also spent 53% more time on site and viewed 23% more pages per visit.

Shopify sees the same thing across its merchants. In Q1 2026, AI-referred sessions converted nearly 50% better than organic search, with order values 14% higher, and AI beat organic search in 23 of 25 merchant categories.

Where they land

The more telling number is where those visitors arrive. Shero Commerce looked at 96.4 million sessions across 73 stores: 51% of AI visits started on a product page, and those visits converted at 1.77%. AI visits that started on the homepage converted at 0.57%.

That's an agency looking at its own clients, so treat the exact figures with care. The direction matches how these visitors behave.

Why they convert

A shopper who arrives from an assistant has usually done the comparing already. They asked a specific question ("trail shoes for mud, wide feet, under $150"), got a short answer with a reason attached, and clicked through to check it. They aren't browsing. They're confirming.

So the product page has one job: confirm what the assistant said, and answer whatever is left.

• Does it fit? Sizing, fit notes, dimensions.
• Does it do what I asked about? The use case, in plain words.
• When will it arrive, and can I send it back? Shipping and returns on the page, not three clicks away.

A page that leaves any of those to guesswork sends a ready buyer back to the assistant, which is happy to suggest someone else.

The same data does both jobs

The details that close the sale are the details that got the product recommended: what it's for, who it fits, what it's made of, and what it takes to get it. An assistant can only recommend what it can read, and a shopper can only buy what the page confirms.

That's why we start with product data. It works at the moment of recommendation and at the moment of purchase.

What the numbers don't say

These are averages across very different stores. AI traffic is still a small share of most stores' visits. Analytics miss shoppers who research in an assistant and come back later by typing your URL. And conversion rates move as the assistants change how they shop. The trend is clear; any single store's numbers will differ.

A check you can run today

1. In your analytics, filter sessions referred by chatgpt.com, perplexity.ai, and gemini.google.com.
2. List the pages those sessions land on.
3. Open your top five landing pages and ask: could a shopper confirm fit, use, shipping, and returns without leaving this page?