Every AI Shopping Channel Outside Amazon Just Went Live

Last week, while most Amazon brand teams were digesting their first Sponsored Prompts bill, something bigger happened outside the walls.

Shopify activated Agentic Storefronts for all 5.6 million merchants. Default on. No setup required. Every product in those stores is now discoverable and purchasable inside ChatGPT, Microsoft Copilot, Google AI Mode, and Gemini. AI-referred traffic to Shopify is up 7x since January 2025. AI-attributed orders are up 11x.

The same week, Gap became the first major fashion brand to enable checkout directly in Google Gemini. OpenAI killed its own Instant Checkout — peaked at 30 merchants out of a promised million — and pivoted to letting retailers build apps inside ChatGPT. Google's Universal Commerce Protocol got multi-item cart support.

Shoptalk ran March 24-26. The executives on stage weren't presenting roadmaps. They were presenting live implementations.

For Amazon brand managers, this isn't a trend piece. It's a competitive intelligence update. Your competitors now have AI distribution channels you don't.

What Actually Happened March 24-26

Five things, same week:

Shopify Agentic Storefronts went default-on. ChatGPT has 880 million monthly active users. Every Shopify merchant — including brands that deliberately chose not to sell on Amazon — is now discoverable when those users ask shopping questions. No listing fees. No FBA required. No advertising spend to surface. The agent finds the product, shows it to the shopper, and handles checkout on the merchant's site.

Gap enabled Google Gemini checkout via UCP. Gap, Old Navy, Banana Republic, Athleta — all live. Product data fed directly to Google (not scraped). They also launched something called an Agentic Sizing Protocol with Bold Metrics: AI-native fit guidance inside the conversation itself. If that sounds small, consider that fit uncertainty is the number one driver of fashion returns. Solving sizing inside discovery changes conversion math.

OpenAI killed Instant Checkout. The model where OpenAI processed the transaction died. Peaked at 30 merchants. The replacement: retailers build dedicated apps inside ChatGPT. Target, Sephora, Nordstrom, Home Depot, Best Buy, Wayfair are all building them. Sephora announced its ChatGPT app at Shoptalk. The AI discovers. The retailer owns the checkout.

Google UCP got multi-item cart and real-time catalog. Cart support was the missing piece — UCP can now handle full shopping sessions, not just single-item purchases. Real-time catalog sync means live pricing and inventory. Endorsed by Walmart, Target, Etsy, American Express, Mastercard, Stripe, Visa. Shopify co-developed it.

Shoptalk confirmed the phase transition. Retail Brew, eMarketer, and Ecommerce Fastlane all covered the same theme: the conversation shifted from "when will AI shopping matter" to "here's what we built." Accenture backed DaVinci Commerce for enterprise AI storefronts — Nestle, Diageo, Nordstrom as clients.

Why This Matters If You Sell on Amazon

The brands presenting at Shoptalk are your competitors. Some of them sell on Amazon too. Some don't. All of them now have AI shopping distribution that works outside Amazon's ecosystem.

A DTC brand on Shopify doesn't need to list on Amazon to show up when ChatGPT's 880 million users ask "what's the best [your product category]." They just need to be on Shopify. Which they already are.

That changes the competitive math for Amazon brand managers. Your Rufus optimization strategy — the listing content, the review signals, the Sponsored Prompts spend — covers one AI shopping surface. There are now at least four others that are live and growing:

Amazon Rufus is still the biggest. Roughly 274 million queries per day, 3.5x conversion rate versus standard search, $12 billion in attributed sales in 2025. That's not changing tomorrow.

But the channels outside Amazon are growing at 7x year-over-year. And unlike Rufus — where you need to optimize listings, manage reviews, and pay for Sponsored Prompts — several of these channels require zero incremental spend. If your competitor is on Shopify, they're already in ChatGPT. You spent money to get into Rufus. They showed up in ChatGPT for free.

The Model That Won

Every surviving AI shopping model routes checkout back to the merchant. OpenAI tried to own the transaction and failed. What works: the AI handles discovery and recommendation. The retailer handles checkout and fulfillment.

This matters for Amazon brands because it means these channels are sticky. Retailers aren't giving up customer data or margins. Shopify keeps its merchant fees and nothing else. Google's UCP is an open protocol. Nobody's asking brands to choose between their own checkout and the AI channel — they get both.

That makes adoption fast. There's no cost to the retailer, no margin hit, and no loss of customer relationship. When adoption costs are zero, distribution scales fast. That's what happened last week.

What Amazon Brand Managers Should Do

First: check your off-Amazon visibility. Ask ChatGPT a shopping question in your category. Ask Google Gemini the same question. See who shows up. If your competitors are there and you're not, that's the gap.

Second: separate your AI shopping strategy from your Amazon strategy. Rufus optimization is not AI shopping optimization. It's Amazon AI shopping optimization. If your brand also sells DTC (via Shopify, your own site, or both), check whether Agentic Storefronts are active for your store. If you're Shopify-based, they should be on by default.

Third: track your share of AI recommendations across platforms, not just Rufus. The percentage of shopping queries handled by AI assistants is growing across every platform. Rufus handles about 14% of Amazon searches today, projected toward 35%. ChatGPT handles an unknown but rapidly growing share of general product research. The brands that track visibility across all of these surfaces will see the full picture. The ones that only watch Rufus won't.

That's what AgentBuy tracks — what AI shopping assistants say about your brand, across platforms, over time. Because the room has five screens now. And most brand teams are still watching one.