Are Omnilux's '3.2% of revenue from AI channels' and Cozy Earth's '20x YoY AI-channel growth' numbers typical, or are those the exceptions Shopify chose to showcase?

What exactly did Shopify publish about Omnilux and Cozy Earth?

Shopify published one sentence naming both brands, in two clauses, in its Spring '26 Edition post of 17 June 2026 (Shopify, June 2026):

• "In March, red light therapy mask brand Omnilux saw AI channels drive 3.2% of total revenue"
• "according to luxury bedding brand Cozy Earth, overall revenue from AI channels was up 20x YoY"

Shopify built the two clauses differently. Shopify states Omnilux's 3.2% as its own figure. Shopify prefaces Cozy Earth's 20x year-over-year rise with "according to," so that number is the merchant's claim, not platform data. The sentence sits under the post's new admin view for managing and reporting on AI channels, offered as proof that "it's already adding up." A showcase figure is the easiest number to drop into a timeline for agentic commerce at a $20M to $100M brand, which is why its exact wording matters.

Why is neither figure a base rate?

Neither Omnilux's 3.2% nor Cozy Earth's 20x is a base rate, because Shopify chose both merchants for a product announcement, and Omnilux works with a paid AI-visibility vendor.

Shopify chose the examples. Shopify launched agentic storefronts, which sell a store's products inside AI chats, on 10 December 2025 (Shopify, December 2025). The June 2026 post added one admin view to manage those AI channels and track their sales, and named two merchants to show it paying off. Picking strong examples is what a launch post is for, so its named merchants are not averages.

Omnilux works at AI visibility. Profound, an AI-visibility software vendor, publishes Omnilux as a customer whose "AI-attributed revenue has grown from roughly 1% to 3% of total revenue" while working with Profound (Profound, October 2026). A brand that pays to track and improve its AI visibility sits at the top of the range, not the middle, in Deploi's judgment.

What does Cozy Earth's 20x tell you without a base?

Cozy Earth's 20x tells you its AI channel grew, and nothing about its size, because Shopify published no starting share. Run two starting points through the same multiple (Deploi arithmetic):

• If AI channels were 0.05% of Cozy Earth's revenue a year earlier, and total revenue held flat, 20x lands at 1%.
• If AI channels were 0.005%, and total revenue held flat, 20x lands at 0.1%.

Both fit Shopify's sentence on Cozy Earth, and they describe two very different businesses. Shopify's own platform multiples carry the same gap: AI referral sessions grew "more than 8x year-over-year" in Q1 2026 (Shopify, May 2026), and AI-driven traffic and orders tripled year over year in Q2 2026 (TechCrunch, August 2026). Neither came with a denominator, which means neither can size a channel for your brand.

Does Omnilux's category explain its 3.2%?

In Deploi's judgment, Omnilux's category helps explain its 3.2%, but Shopify publishes no AI revenue share by category. Omnilux sells red light therapy masks (Shopify, June 2026), which we read as a researched, specification-heavy skincare purchase. Shopify's guide says Perplexity Shopping "tends to save shoppers the most time when making research-heavy purchases in categories like electronics, home goods, and skin care" (Shopify blog, April 2026). Shopify's Q1 2026 data says only that AI-referred conversion beat organic search on product page sessions in 23 of 25 merchant categories, without naming them (Shopify, May 2026). Shopify also said 75% of AI-attributed purchases in Q2 2026 happened outside its top 100 categories (TechCrunch, August 2026). So AI sales spread across many categories, and one skincare brand's March is weak evidence for a bedding, apparel or hardware brand.

What do typical stores measure instead?

Typical stores measure AI at a small fraction of Omnilux's 3.2%, judging by two independent samples: 0.12% of revenue across 10 US Shopify stores, and 0.26% of orders across 73 stores, 42 on Shopify (Ethercycle and Shero Commerce, September 2026).

The gap between 3.2% and 0.12% is about 27 times (Deploi arithmetic). Both numbers can be true, of different merchants, categories, months and methods, because, in Deploi's judgment, the variance between brands is far larger than the trend. Ethercycle calls its AI-optimized sample "closer to a ceiling than a floor," so a store with no AI work may sit lower still.

How should you use a showcase number?

Use a showcase number as a ceiling and a category test, never as a forecast input.

• As a ceiling. Omnilux's 3.2% of March 2026 revenue (Shopify, June 2026) is one well-suited brand's month, backed by paid AI-visibility work with Profound (Profound, October 2026). Treat it as the top of the range for a brand like that, not the middle of yours.
• As a category test. In Deploi's judgment, specification-heavy, comparison-shopped products sit nearer the Omnilux end, and impulse or fashion-led products nearer the 0.12% of revenue that 10 US Shopify stores drew from AI (Ethercycle, September 2026).
• Never as a forecast input. A board model assuming 3% of revenue from AI next year assumes the showcase case, off a base nobody published. Start from your own admin data, then see how to budget AI-search work.