Amazon still leads in online sales, but a newly launched AI Commerce ranking puts it far down the list — No. 158 — behind many much smaller merchants. That gap is not a measurement error: it exposes a practical shift in how shoppers find products when AI agents — third‑party systems that surface and recommend items — are part of the discovery path.

What the AI Commerce rankings actually measure

Digital Commerce 360 and ReFiBuy’s AI Commerce Rankings assess Top 1000 online retailers on how effectively they appear through external AI agents. Rather than tracking sales, the index evaluates observable signals such as catalog accessibility to third‑party agents, volume of traffic from those AI sources, the diversity of agent channels sending traffic, and short‑term momentum over 90‑day windows.

The result is a leaderboard that diverges from conventional sales lists: Amazon — the No. 1 merchant by ecommerce sales in the Top 1000 — sits at No. 158 in the AI index, while Walmart (No. 2 by sales) is No. 37. That wide gap shows scale alone does not guarantee prominence when discovery is mediated by external AI platforms.

Why Amazon falls behind on AI-driven discovery

The rankings indicate Amazon has intentionally limited how third‑party AI agents access its product metadata and checkout flows. Scot Wingo, ReFiBuy’s founder and CEO, described Amazon as having “closed all four doors” — meaning the channels the index tracks are largely unavailable. This is consistent with Amazon’s broader strategy: it operates Amazon Web Services (AWS) and its own AI experiences (like Alexa for Shopping and Buy for Me), creating incentives to keep shoppers inside proprietary agentic paths.

Put another way: AI agents reward structured, machine‑readable product data that they can index and surface. A platform that restricts that access will perform poorly on rankings designed around reach through external agents, even if its sales volume is enormous.

Smaller merchants are getting disproportionate visibility

The AI Commerce data shows smaller retailers can punch above their weight when they make catalogs accessible and agent‑friendly. In Q2 2026, Nixon, Online Labels and Everlane ranked first through third in the AI index while sitting at No. 722, No. 814 and No. 264 in sales‑based Top 1000 standings. That pattern suggests third‑party AI systems such as ChatGPT and Gemini can channel meaningful shopper attention to merchants that adopt agent‑friendly data practices.

Category structure matters too. Automotive Parts & Accessories — a segment with large, structured catalogs — was flagged as especially well positioned for AI discovery if merchants improve bot friendliness and catalog detail. AI agents favor granular, standardized attributes that let them match shopper intent to the right SKUs.

Practical implications for marketers and ecommerce teams

The divergence between sales rank and AI rank creates a clear operational choice: invest in catalog readiness for agentic discovery or accept that some referral channels will be limited. Three concrete implications follow:

  • Make product data agent‑friendly. Publish structured feeds and machine‑readable attributes — rich product descriptions, standardized identifiers (GTINs), fitment and size metadata — in formats AI developers expect. This is about discoverability, not exposing sensitive transactional systems.
  • Test third‑party AI distribution. Allow indexing by major AI platforms where commercially sensible and measure the incremental traffic and conversions. The AI index shows that openness can deliver referral volume that smaller merchants would otherwise lack.
  • Factor platform incentives into strategy. Large retailers that run their own AI products or cloud services may favor keeping interactions inside their ecosystems. Plan partnerships and technical integrations with that reality in mind, and negotiate commercial terms if you rely on a platform’s catalog access.

These actions are structural rather than purely technical: improving catalog completeness, adopting agent‑friendly schemas and formalizing distribution agreements are direct levers the AI Commerce Rankings track.

What to watch next

Two trends will determine whether AI referrals become a lasting channel. First, merchants that prioritize standardized, enriched product metadata are likely to gain referral traffic faster. Second, platform owners will keep balancing openness against the commercial value of retaining users within their AI experiences. How those trade‑offs settle will shape which merchants can rely on third‑party agents for discovery.

For ecommerce leaders the operational next step is straightforward: audit your product feeds for completeness and schema consistency, prioritize machine‑readable attributes, and run controlled tests with major AI platforms. Those steps will show whether agentic channels convert into customers for your business — and whether catalog openness can be a cost‑effective path to discovery where scale alone no longer guarantees visibility.