Agentic Commerce Has Arrived: A Brand's Playbook for Selling Inside ChatGPT and Gemini

Etsy stock jumped 16% the day OpenAI announced that ChatGPT users could buy products without leaving the chat. Google followed weeks later with checkout inside Gemini, launching Walmart as the first partner. The conversion moment is moving out of your storefront and into chat windows you don't control.

Most coverage focuses on how the protocols work. This post focuses on what happens next: the shift in your site's role, the post-purchase gap that opens up, and the playbook for brands and businesses that want to own the ecommerce customer relationship after an agentic sale.

What is agentic commerce? In short, it's the shift from humans clicking "buy" to AI agents completing entire purchase flows on a shopper's behalf. The agent handles discovery, comparison, payment, and order confirmation inside a single conversation. OpenAI's Agentic Commerce Protocol, Google's Universal Commerce Protocol, and Amazon's closed-loop Rufus system are the three architectures powering this shift. For a full primer, see our complete guide to agentic commerce.

Three Eras, One Trend: The Conversion Surface Keeps Moving

From 2016 to 2022, chatbots linked shoppers to product pages. Your site was the storefront and the checkout. Attribution was clean.

From 2023 to 2025, generative AI and RAG-powered assistants like Alhena AI's Shopping Assistant could make product recommendations for the right SKU and populate carts inside a conversation. But the customer still checked out on the brand's domain.

In 2026, the third era arrived. OpenAI's Agentic Commerce Protocol and Google's Universal Commerce Protocol let Autonomous AI agents hold the cart, process payments through secure token technology to autonomously handle transactions. This orchestration lets autonomous agents complete tasks, and hand orders directly to the merchant. Major retailers and businesses including Shopify merchants, U.S. Etsy sellers, Walmart, Target, and other global retailers are already connected through native integrations and APIs. The brand site is no longer the conversion surface.

It’s a change that affects key dynamics about attribution, CAC, and first-party data collection.

Where Each Platform Fits Your Catalog

We've published a full comparison of ChatGPT, Gemini, and Perplexity for merchants. Here’s the short version:

• ChatGPT: Task-completion intent. Users ask specific questions and expect a short list with a buy button. Strongest for considered purchasing decisions on Shopify and Etsy marketplaces.
• Gemini: Consumer research-heavy intent tied to Google's Shopping Graph. Best for consumables, replenishment, and retailers already in Google Merchant Center.
• Perplexity: High-intent, research-deep consumers. Smaller retailer pool, but growing fast. Best for niche and speciality products.

If you sell across categories, it’s important to note that you need to show up on all three platforms. For the technical details on feed optimization requirements and checkout flows, see our ChatGPT product feed guide and instant purchasing flow breakdown.

Amazon Rufus vs. Gemini vs. ChatGPT: Three Models of the Same Trend

All three platforms let AI agents buy products for shoppers. But they run on fundamentally different architectures, and that difference shapes your strategy.

Closed-loop: Amazon Rufus

Rufus operates entirely inside Amazon's ecosystem. It's trained on Amazon reviews, purchase histories, and product Q&A data rather than the open web. Over 250 million shoppers use Rufus, and customers who engage with it convert 60% better than those who browse traditionally, according to Retail Technology Innovation Hub.

Amazon's "Buy for Me" feature, launched in April 2025, extends Rufus to external merchant sites. The agent visits third-party stores, adds items to cart, and completes checkout, all routed through Amazon's payment and returns infrastructure. In April 2026, Amazon added Scheduled Actions: Rufus can now place recurring orders on a calendar, trigger purchases at a price threshold, or reorder when it predicts a household is running low. For merchants, the tradeoff is clear. Rufus can drive volume, but Amazon owns the customer data, the payment relationship, and the return flow. Brands get sales, not relationships.

Open-protocol: ChatGPT and Gemini

ChatGPT (via ACP) and Gemini (via UCP) take the opposite approach. These agents discover products across the open web and route checkout to merchant-controlled endpoints. Stripe handles payment tokenization for ACP; Google Merchant Center powers UCP. The brand keeps first-party customer data, controls the checkout experience, and owns the post-purchase relationship.

Walmart, one of the first UCP partners, already sees roughly 20% of its referral traffic from ChatGPT. The open-protocol model rewards structured product data, JSON-LD schema, and agent-readable PDPs, not marketplace listing optimization.

What each model rewards

• Rufus rewards Amazon SEO: optimized titles, A+ content, review velocity, and competitive pricing within Amazon's catalog.
• ChatGPT and Gemini reward feed quality, schema.org markup, agent-readable product pages, and protocol enablement. Marketing copy matters less than structured attributes.
• All three hand the shopper back to the brand for post-purchase support. None of them handle returns, order tracking, or product questions after the sale. That gap is where your on-site AI layer, from tools like Alhena's Support Concierge, becomes essential.

If you sell on Amazon and DTC, you're playing both models. Treat them as separate channels with separate optimization strategies. Your Amazon listings win Rufus; your product feeds and structured data win ChatGPT and Gemini. And regardless of which agent closes the sale, the customer lands on your site for everything that happens next.

Google made the announcement official at the National Retail Federation's annual convention in January 2026. CEO Sundar Pichai laid out a vision for AI-assisted commerce where Gemini surfaces products from participating retailers and lets shoppers buy without leaving the chat. Walmart, Shopify, Wayfair, Target, and Etsy signed on as founding UCP partners, with 20+ additional companies endorsing the protocol in the weeks that followed.

The Walmart integration was the first to go live. Shoppers who link their Walmart and Gemini accounts get product recommendations shaped by past purchase history. Items picked through Gemini drop straight into an existing Walmart or Sam's Club cart. John Furner, Walmart's incoming president and CEO, called it "the next great evolution in retail." The feature launched as US-only, with international expansion planned.

Then, in March 2026, Gap became the first major fashion retailer to enable full checkout inside Gemini. Announced at Shoptalk Spring in Las Vegas, the integration lets shoppers browse Gap's catalog, get styling recommendations, and complete a purchase entirely within the Gemini interface, with payment handled through Google Pay and fulfillment managed by Gap. No redirect to the brand's website required.

Gap's launch matters because it proves the UCP pipeline works end-to-end for a real retailer with complex sizing, seasonal inventory, and multi-brand operations. Salesforce estimated that AI influenced $272 billion in global retail sales during the most recent holiday season alone. With live checkout now running inside Gemini for both mass retail (Walmart) and fashion (Gap), the question for mid-market brands isn't whether agentic commerce will reach their category. It's whether their post-checkout experience is ready when it does.

The Real Shift: Your Site Becomes a Post-Purchase Hub

If 10 to 30% of product discovery and the shopping journey move into LLMs over the next 18 months, your e-commerce brand site doesn't disappear. It transforms. The site becomes the fulfilment surface, the post-purchase hub, and the place where brand identity lives.

But here's the gap most brands aren't thinking about: when a customer buys your product inside ChatGPT, they won't go back to ChatGPT for order status, returns, or product questions. They'll come to you. And if your customer experience is slow, scripted, or disconnected from the sale, you lose the relationship you just paid to acquire.

That makes the post-purchase conversation the new battleground for e-commerce customer loyalty, repeat revenue, and long-term retention.

Owning the Post-Checkout Conversation

The businesses winning in agentic commerce aren't just optimising feeds. They're building an AI-powered support layer that picks up exactly where the agentic sale ends.

Alhena's Support Concierge handles this layer: order tracking, returns, exchanges, product guidance, and self-service solutions across Instagram, WhatsApp, email, web chat, and other channels. It deploys in under 48 hours with pre-built integrations with no dev resources or professional services, and every response is grounded in verified product data, personalisation, and full order context for every interaction, so there are no hallucinated answers confusing a customer who just bought from an AI agent.

Brands like Tatcha have driven 3x conversion rates and 82% chat deflection by owning the full post-purchase customer journey through conversational AI. Manawa cut response times from 40 minutes to 1 minute with 80% inquiry automation and support tasks handled by AI. These performance results show how businesses can treat AI support as a revenue channel, not just a cost centre but a key revenue-driving services layer.

A Four-Step Playbook for 2026

1. Enable the protocols. Connect to ACP and UCP through your existing payments infrastructure (Stripe, Mastercard, Visa, or other processors). If you're on Stripe, it can be one line of code to enable payments. See our ACP readiness guide. For how MCP fits alongside ACP and UCP, see our protocol comparison for details.
2. Run product feed optimization. Structured identifiers (GTIN, mpn, brand), machine-readable attributes (ingredients, materials, return policy), and JSON-LD Product schema on your PDPs matter more than marketing copy. Our product feed guide covers the specifics.
3. Open your site to AI agents and crawlers. Review your robots.txt and add an llms.txt file. The pages that survive the AI shift are the ones serving structured, machine-readable data signals.
4. Own the post-purchase experience. Deploy AI systems as a support layer that handles order management, returns, and product guidance across every channel, including web chat and social platforms. This is where you keep the customer relationship and buying and shopping experiences that agentic commerce creates in the digital commerce landscape.

Ready to own the post-checkout conversation? Book a demo with Alhena AI or start for free with 25 conversations.

Alhena AI

Schedule a Demo

Frequently Asked Questions