Shopify Reports AI Purchases Favor Long-Tail Categories
Shopify President Harley Finkelstein reported on the company's August 5 earnings call that 75% of AI-attributed purchases in the quarter came from product categories outside Shopify's top 100. CNBC, TechCrunch and PYMNTS reported that AI-referred traffic and orders to Shopify stores each tripled year over year, indicating that AI-mediated discovery is reaching specialized inventory beyond the platform's largest categories.
Shopify President Harley Finkelstein reported on the company's August 5 earnings call that 75% of AI-attributed purchases during the quarter came from product categories outside Shopify's top 100. CNBC, TechCrunch and PYMNTS reported that traffic and orders referred from AI platforms to Shopify stores each tripled year over year.
Finkelstein cited examples including a car seat designed to fit three across a sedan and reef-safe sunscreen that does not leave white residue, according to PYMNTS' account of the call. The examples illustrate the claim Shopify is making: natural-language discovery can connect shoppers with specialized products that may not lead popularity-ranked search results.
The long tail was already a large market
Shopify's May 11 analysis provides first-party background for the quarterly figures. It defined the long tail as categories outside the top 100 by gross merchandise value and reported that those categories represented nearly 55% of all Shopify sales. Shopify also said 71% of AI-attributed orders in 2025 came from the long tail, nearly 54% of new stores launched in 2025 belonged to long-tail categories, and 41% of stores launched with a single product.
The newer 75% quarterly figure was reported by the three independent outlets from the August earnings call. TechCrunch also reported Finkelstein's statement that traditional search remained one of merchants' largest traffic sources: traditional-search sessions had risen 1.3 times over two years and still accounted for roughly one-third of storefront sessions. His framing was that AI discovery complements conventional search rather than replacing it.
TechCrunch reported another journey-level metric: half of AI-referred sessions landed directly on a product-description page, 2.5 times the rate for traditional search. That can shorten the path from discovery to a specific item, but the available reports do not establish whether the traffic mix itself caused higher conversion.
Product data and attribution become operational constraints
For retail data teams, Shopify's figures make product information quality a practical dependency. A system matching a detailed request needs structured attributes, inventory status, variant identifiers, pricing and merchant policies. If those fields are incomplete or stale, natural-language retrieval can still return a plausible but unsuitable product.
CNBC reported that Shopify's Sidekick assistant handled 34 million conversations in the second quarter. Finkelstein also said merchants using Sidekick during onboarding reached their first five sales faster, although CNBC did not provide a numerical comparison or methodology for that outcome.
The reported 75% share concerns purchases Shopify attributes to an AI-powered discovery path, not every transaction influenced by AI. Shopify's May methodology defines an AI-attributed order as one where the buyer's discovery path included an AI-powered channel, but the August reports do not break the quarterly figure down by model provider, merchant size, category or attribution window. Those omissions matter when teams compare referral quality or decide how much weight to place on a rapidly growing but still narrowly measured channel.
Key Points
- 1Shopify reported that 75% of quarterly AI-attributed purchases came from outside its top 100 categories, while AI-referred traffic and orders each tripled year over year.
- 2Shopify's May first-party analysis reported that long-tail categories generated nearly 55% of platform sales and 71% of AI-attributed orders in 2025.
- 3TechCrunch reported that half of AI-referred sessions landed directly on a product page, 2.5 times the rate for traditional search, while the available reports leave attribution methodology and category breakdowns unresolved.
Scoring Rationale
The reported metrics provide a specific view of AI-assisted discovery on a major e-commerce platform, including its reach into specialized product categories. The story is relevant to teams building retail search, catalog infrastructure, attribution systems and agentic-commerce integrations, but the quarterly reporting does not disclose provider, category or attribution-method detail.
Sources
Public references used for this report.
Practice with real Retail & eCommerce data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Retail & eCommerce problems
