The AI Commerce Rankings Are In: Small Retailers Are Winning the Agent Era

There is a retailer sitting at number 814 in Digital Commerce 360's Top 1000 by estimated web sales. In the new AI Commerce Rankings for Q2 2026, it sits at number one.

That is the headline finding from the first quarterly results of a benchmark that is already making brand directors uncomfortable. Digital Commerce 360 and ReFiBuy launched the AI Commerce Rankings in July as a co-developed index measuring how prepared the Top 1000 retailers are for agentic shopping. The Q2 data dropped on August 20, and the results scramble the conventional pecking order almost completely.

Catalogue Quality Beats Catalogue Size

The rankings score every retailer across four signals: bot friendliness, AI-source traffic, diversity of AI sources, and 90-day momentum. Sales history is not in the formula at all. The methodology reflects a hard commercial reality: AI shopping agents do not care how big your Black Friday was. They care whether your product data is machine-readable, complete, and pushable across the platforms where agents are sourcing right now: ChatGPT, Google Gemini, Alexa for Shopping.

The top five retailers in the Q2 AI rankings include Nixon (watches), Online Labels, and CustomInk. Those three companies sit at positions 722, 814, and 264 respectively in the traditional Top 1000 by web sales. Nixon, the watch brand, outranked the vast majority of household-name retailers on the metric that will increasingly determine who gets recommended at checkout.

Across merchandise categories, Automotive Parts and Accessories retailers climbed an average of 20.1 positions in the AI rankings during Q2. Flowers and Gifts fell by an average of 17.8 points. Lawn and garden, tools, and fishing and hunting sellers posted some of the strongest positive movement at the subcategory level. Beer, wine and spirits, books, and musical instrument sellers saw the steepest declines. That spread is not random: the winning categories tend to sell high-specificity, spec-driven products where a structured product feed gives an AI agent the exact attributes it needs to confidently make a recommendation.

Why This Matters Right Now

Traffic from AI platforms to US retail sites was up 393% year over year in Q1 2026, according to Adobe Analytics data cited by ReFiBuy. In May alone, AI-platform traffic to online retailers more than doubled year over year, growing 138%, per Adobe Analytics figures quoted by Digital Commerce 360. The growth curve is steep enough that retailers who delay catalogue optimisation are not just missing upside; they are actively ceding ground to smaller operators who moved earlier.

Agentic commerce is not a 2028 planning assumption. Shoppers are already starting product searches inside ChatGPT and Google AI overviews, and autonomous agents are completing checkouts on their behalf today. McKinsey has pegged the potential scale of agentic retail at somewhere between $900 billion and $1 trillion in US revenue by 2030.

Amazon also moved quickly this week. Starting August 20, it brought its Warehousing and Distribution service to Germany, France, Italy, Spain, and the UK, giving sellers long-term bulk storage in distribution centres with automated replenishment to FBA centres across Europe. For DTC brands using Fulfilment by Amazon, that is a material operational unlock on the continent, compressing the gap between demand signals and shelf availability.

What the Trademark Data Tells You

Here is the part that should worry any brand founder who has been sitting on their IP paperwork. Class 3 UK trademark filings, the class covering cosmetics, skincare, and personal care, came in at just 1,845 in Q3 2026, down 37.2% against the prior period, per AIBD analysis of IPO (TMD) data as of August 2026. That is a dramatic contraction at precisely the moment when beauty and personal care DTC brands are supposedly racing to establish defensible brand positions before AI agents lock in their category recommendations.

The two trends connect directly. An AI shopping agent surfacing skincare recommendations will default to brands with established, structured, credible identities. A brand without a registered trademark has one fewer signal of legitimacy to feed into that agent's decision process. The operators winning the AI Commerce Rankings right now did not get there by accident: they built complete, coherent brand and product data infrastructure. Trademark registration is part of that stack.

The Operator Playbook

So what do you actually do with this? Three things, roughly in order of urgency.

First, run your catalogue through an agentic readiness audit before Q3 ends. The AI Commerce Rankings score on bot friendliness and AI-source diversity; both are addressable with structured data work, not a full platform rebuild. ReFiBuy has made the methodology public enough to run a rough internal diagnostic.

Second, stop thinking about search engine optimisation and start thinking about agent optimisation. The brands rising in the AI rankings are the ones whose product data is accessible, interpretable, and specific. Long-tail attribute completeness, the stuff that felt like SEO housekeeping two years ago, is now a direct revenue driver.

Third, sort the trademark. Class 3 filings are down hard. The queue at the IPO is shorter than it has been in years. Counterintuitively, that is an opportunity: fewer competing applications means faster examination timelines. The AI agent era rewards brands that can prove they exist.