Skip to main content
Nova Analytics
Back to News
Update
AI & Technology

AI shoppers arrive ready to buy but rarely close the sale online

August 10, 2026
5 min
Summarize with AI
ClaudeClaudeGeminiGeminiChatGPTChatGPT
A

CEO at Nova Analytics

LinkedIn

Antoine founded Nova Analytics to empower Amazon sellers with enterprise-grade analytics. He specializes in data architecture and building scalable solutions for e-commerce businesses.

Quick Summary

  • •Etsy: AI agent platforms send under 1% of traffic but convert higher with bigger average orders
  • •Adobe: AI-referred traffic up 393% year over year, revenue per visit 37% higher as of March
  • •48% of shoppers researched their last purchase with AI; only 49% would let an agent complete it
  • •Trust falls to 3% among consumers who are not already regular AI users
  • •For Amazon sellers the lever is listing data quality, not a new traffic channel

Nova surfaces every Amazon fee, refund, and margin shift in your live P&L, across 21 marketplaces. Open the live P&L

What happened

Second-quarter earnings calls have produced the first consistent picture of what AI assistants actually do for retailers. The traffic is excellent and there is almost none of it. Etsy told analysts on August 6, 2026 that AI agent platforms send under 1% of its total traffic, while converting at a higher rate with a larger average order size. DoorDash reported the same shape: agentic order volume from AI partners remains low (PYMNTS, August 7, 2026).

The quality side is measurable. Adobe data covering more than a trillion visits to US retail sites put AI-referred traffic growth at 393% year over year in the first quarter of 2026, with revenue per visit 37% higher than non-AI traffic as of March, 48% longer time on page and 13% more pages viewed.

Where the funnel breaks

At the checkout. PYMNTS Intelligence found 48% of online shoppers used AI to research their most recent purchase, but only 49% would let an agent complete the transaction, and that falls to 3% among people who do not already use AI regularly. Research is delegated. Payment is not.

Shopify's read on the same quarter is worth pairing with it: AI search is adding traffic on top of Google rather than replacing it (TechCrunch, August 5, 2026). Nothing here says the old channels are going away.

What it means for Amazon sellers

Amazon's assistant sits inside the buying flow rather than referring traffic to your listing, so you will never see an AI line in your reports the way a Shopify merchant might. The signal still matters, because the thing that gets a product picked by an assistant is the same thing that gets it picked in filtered search: accurate attributes, honest titles, complete specifications and reviews that describe real use.

The advice that follows is unglamorous. Assistants read structured data before they read your marketing copy. A listing with the wrong material, missing dimensions or a mismatched pack size loses to a duller listing that is complete.

What to change in the next 72 hours

  1. Audit attributes on your top 20 ASINs. Material, dimensions, pack quantity, compatibility. Fill every field that applies, even the ones no shopper reads.
  2. Fix contradictions between title, bullets and the attribute table. A conflict is worse than a gap, because it makes the whole record less trustworthy to a machine reader.
  3. Do not build an AI channel strategy on under 1% of traffic. Treat it as a reason to fix data quality, not as a reason to reallocate budget.
  4. Keep margin per SKU current. When a new source of volume does arrive, you want to judge it on contribution rather than on session counts.

How Nova helps

  • Listing Health Scanner - surfaces incomplete and inconsistent listing data across your catalogue, which is exactly what machine readers penalise.
  • Nova MCP - connect your Amazon data to Claude, ChatGPT and other assistants so you can question your own numbers in plain language.

Frequently Asked Questions

Common questions about this topic

Small in volume, strong in quality. Etsy told analysts on August 6, 2026 that traffic from AI agent platforms remains under 1% of its total, though it converts at a higher rate with a larger average order size. DoorDash said agentic order volume from AI partners stays low across the board.
Adobe data covering more than a trillion visits to US retail sites put AI-referred traffic growth at 393% year over year in the first quarter of 2026, with revenue per visit running 37% above non-AI traffic as of March. Those visitors also spent 48% longer on the page and viewed 13% more pages.
At the purchase step. PYMNTS Intelligence found 48% of online shoppers used AI to research their most recent purchase, but only 49% would let an AI agent complete the transaction, and trust drops to 3% among consumers who do not already use AI regularly.
Amazon's own assistant sits inside the buying flow rather than referring traffic to you, so you will not see an AI channel in your reports. The practical read is that AI-assisted research is now part of the path to purchase, which raises the value of accurate titles, bullets and attributes over clever copy.
Not as a channel. Treat it as a measurement problem: know your margin per SKU well enough that any new source of volume can be judged on contribution rather than on traffic counts.

Verified Sources

All information verified from official Amazon sources and trusted industry analysts as of publication date.