Small Retailers Appear in Just 1–4% of AI Shopping Recommendations, Lightspeed Study Finds

Independent and local retailers appear in just 1 to 4 percent of AI-generated shopping recommendations, according to a Lightspeed Commerce study of 460,000 AI shopping responses published October 2. That near-invisibility signals a structural threat to small retail — one driven not by price competition, but by algorithmic exclusion.

When a shopper asks an AI assistant where to buy hiking boots, a kitchen appliance, or a birthday gift, the answer almost always names a major chain. Not because independent retailers are worse — in many cases they offer better selection, more knowledgeable staff, and competitive prices — but because the data infrastructure that makes a retailer legible to an AI favors businesses that invested in it years ago.

What Lightspeed Found

Lightspeed, which provides point-of-sale and e-commerce software to independent retailers globally, analyzed 460,000 AI-generated shopping recommendations across ChatGPT Shopping and Google AI Mode. The results were stark:

• 97% of recommended retailers were large or mid-sized chains
• 1–4% combined share for independent and local merchants
• The gap held across product categories: apparel, electronics, home goods, and specialty items
• Amazon, Walmart, Target, and Best Buy dominated recommendation surfaces

The study draws from actual AI responses at scale — not simulated queries — giving it empirical weight that anecdotal complaints about AI search have lacked.

Why AI Systems Prefer Big Chains

This is not a deliberate design decision against small business. It is an emergent consequence of how AI shopping systems learn and operate.

Large retailers have spent years building structured data infrastructure: Google Merchant Center product feeds, schema.org markup on product pages, inventory APIs, and direct integrations with shopping platforms. When an AI system parses the web to answer a product query, it finds chains' data clean, consistent, and immediately actionable.

Most independent retailers have not made these investments. Their product listings may lack structured markup. Their review profiles may be thin. Their return policies may not be programmatically readable. From an AI's perspective, recommending a boutique with 40 reviews introduces uncertainty that recommending a national chain does not.

AI shopping models also optimize for what might be called transactional confidence — the probability that the recommendation will satisfy the user without friction. A retailer with a known return policy, reliable shipping, and thousands of verified reviews wins that calculation regardless of whether a smaller competitor is genuinely superior on price, service, or expertise.

The Agentic Commerce Problem

The timing matters because AI shopping is becoming more autonomous, not less.

Tools like OpenAI's Operator and Google's Astra are designed to handle purchases on behalf of users — researching products, selecting a retailer, and completing the transaction without the user browsing independently. In that model, there are no page-two search results, no Google Shopping ads that a small retailer can purchase for visibility, and no long tail of results where a niche merchant might rank.

If the AI does not name you in its first recommendation, you do not exist in the transaction.

For the past decade, small retailers adapted to each shift in online discovery — from directory listings to SEO to Google Shopping to social commerce. Agentic AI compresses those cycles dramatically. A merchant that is not AI-legible today faces a harder path each month as AI assistants capture more of the purchase journey upstream.

The Lightspeed study frames this as an existential issue for the long-tail retail economy — the independent bookshops, specialty kitchen stores, local running shops, and boutique clothiers that are not competing on Amazon-scale logistics but on local expertise, curation, and community. None of those advantages are legible to an AI shopping model.

Industry Response and Policy Pressure

Small business advocates are calling for two specific interventions:

1. Disclosure requirements — AI shopping assistants should reveal the criteria used to select and rank retailers, similar to ad disclosure requirements in search
2. Fair-ranking standards — analogous to what the EU's Digital Markets Act imposes on app store operators, applied to AI commerce channels

The Lightspeed study arrives as the FTC is conducting an ongoing inquiry into AI platforms and competition. Whether that inquiry produces enforceable action quickly is uncertain. What is not uncertain is that the traffic shift is already underway — retailers waiting for regulatory protection before adapting are already losing ground.

Some technology vendors are developing tools to help independent merchants build AI-readable product data infrastructure. Lightspeed's publication of this study is, in part, a commercial argument for its own services — but the data stands independent of the motivation.

The practical path forward for small retailers: invest in structured product data now. Build out Google Merchant Center feeds. Generate reviews systematically. Make inventory machine-readable. It is unsexy infrastructure work, but it is table stakes for AI-era retail visibility.

What to Watch

Congressional testimony and FTC filings citing the Lightspeed data are likely within 90 days — small business associations have been searching for empirical grounding to argue that AI commerce platforms require new rules. Watch also for whether Google or OpenAI introduce voluntary diversity-in-recommendation policies, which would be easier to implement than legislation but harder to enforce independently.

For retailers: the window before agentic shopping becomes mainstream is narrowing. Treating structured data investment as optional is no longer a defensible posture.

Sources: Lightspeed Commerce analysis via Fortune