Agentic Commerce: The 2026 Guide for Retailers

The process typically unfolds in four stages:

1. Intent Capture.The consumer describes what they want in natural language. “I need a birthday gift for my sister, she likes skincare, budget $75.” The AI agent interprets this intent, extracting product category, price constraints, recipient preferences, and urgency.

2. Discovery and Matching. The agent queries merchant product catalogs using structured data feeds. Unlike traditional search engines that rank web pages, AI agents evaluate product attributes directly — price, availability, shipping speed, reviews, and relevance to the stated intent. Merchants whose catalogs are optimized for machine readability get recommended. Those with incomplete or unstructured data get skipped.

3. Evaluation and Recommendation.The agent compares options across multiple merchants, weighing factors the consumer specified. It presents a curated shortlist with reasoning: “This moisturizer from Brand X has 4.8 stars, ships free in 2 days, and fits your budget at $68.” The consumer reviews and approves.

4. Transaction Execution. The agent completes the purchase using stored payment credentials, handles shipping selection, and confirms the order — all within the chat interface. Some platforms support fully autonomous purchasing for low-risk items, while higher-value transactions require explicit consumer approval.

This four-stage flow is enabled by a growing ecosystem of open protocols. The Agentic Commerce Protocol (ACP), developed by Stripe and OpenAI, powers ChatGPT Shopping product discovery and merchant redirect. Google's Universal Commerce Protocol (UCP), launched at NRF 2026, covers the full commerce journey from discovery to post-purchase. Anthropic's Model Context Protocol (MCP) provides the data connectivity layer that lets agents access real-time inventory, pricing, and product details.

Consumer adoption is accelerating faster than mobile commerce did. According to a Morgan Stanley AlphaWise survey of 2,000 U.S. consumers (Nov 2025), 45% had used ChatGPT in the prior month. A January 2026 IBM Institute for Business Value study found 45% of consumers already use AI for some part of the buying journey. And 38% of consumers have used generative AI for online shopping, per Adobe (2025).

Platform infrastructure is production-ready. ChatGPT Shopping, which evolved from early shopping integrations in late 2025, is now live for all U.S. users, recommending products and directing shoppers to merchant sites. Etsy and over 1 million Shopify merchants are live. Google launched AI Mode with agentic checkout capabilities, with Wayfair, Chewy, and Etsy among early participants. Microsoft Copilot supports checkout via ACP. Perplexity offers instant purchasing through its PayPal-powered shopping feature.

Open protocol standards are reducing integration complexity. Before September 2025, connecting to AI shopping channels meant custom integration for each platform. The launch of ACP (Sep 2025), followed by MCP's rapid enterprise adoption, and UCP's debut at NRF 2026 (Jan 2026), has created a standardized foundation.

The financial stakes are significant. Adobe Analytics data shows AI-referred visitors have a 38% higher purchase completion rate compared to traditional search visitors (Adobe, Black Friday 2025). Retailers with AI agent integration saw roughly 7x better sales growth during Cyber Week 2025 than those without (Salesforce data, Dec 2025).

November 2024. Anthropic launches the Model Context Protocol (MCP), the first open standard for connecting AI agents to external systems and data. MCP later becomes the foundation of every major commerce protocol that follows.

April 2025. Google launches Agent2Agent (A2A), the first interoperability protocol for AI agents to coordinate across systems. Donated to the Linux Foundation by year-end.

September 2025. Stripe and OpenAI release the Agentic Commerce Protocol (ACP) under Apache 2.0. ChatGPT Shopping launches with Etsy, Glossier, SKIMS, Spanx, and Vuori. Google announces AP2 the same month for secure agentic payments.

October-November 2025. Adobe Analytics records 805% YoY growth in AI-driven retail traffic during Black Friday 2025. Salesforce reports $67 billion in AI-influenced Cyber Week sales. Retailers with AI agent integrations see 7x sales growth versus those without.

January 2026. Google CEO Sundar Pichai unveils the Universal Commerce Protocol (UCP) at NRF 2026, co-developed with Shopify, Etsy, Wayfair, and Target. UCP covers the full commerce journey from discovery through post-purchase, with 20+ payment and retail partners.

February-March 2026. Mastercard completes its first live agentic transactions in Singapore, South Korea, and other Asia-Pacific markets. Visa launches its Trusted Agent Protocol commercially after sandbox piloting with 100+ partners. The major card networks begin supporting agent-initiated payments at scale.

April 2026. American Express debuts Agentic Commerce Experiences (ACE) Developer Kit and announces industry-first purchase protection for registered AI agent purchases. With Mastercard already live and Visa TAP commercial, all three major U.S. card networks now support agentic commerce.

March 2026. OpenAI deprecates ChatGPT Instant Checkout. The model shifts from "buy in the chat" to "discover in the chat, transact on the merchant site." Brands keep customer relationships, login data, and loyalty engagement.

April 2026. Microsoft adopts UCP in Merchant Center and wires Shopify Catalog into Copilot. Combined with ChatGPT on ACP, mid-market retailers now feed two protocols. Adobe Summit 2026 ships an MCP server for Adobe Commerce. The protocol layer is no longer optional.

ChatGPT Shopping (ACP)

Etsy, Glossier, SKIMS, Spanx, Vuori, and over 1 million Shopify merchants are live on ChatGPT Shopping, with Walmart and Target announced as upcoming partners, powered by the Agentic Commerce Protocol and Stripe. OpenAI reports that Product recommendations are based on relevance, not enrollment status (OpenAI, Feb 2026).

Stripe Agentic Commerce Suite

Stripe's Agentic Commerce Suite, launched December 2025, enables merchants to sell through any ACP-compatible AI agent. URBN (Anthropologie, Free People, Urban Outfitters), Coach, Kate Spade, Revolve, Ashley Furniture, and Etsy are live (Stripe, Dec 2025).

Google AI Mode and UCP

Google launched the Universal Commerce Protocol at NRF 2026 with co-development from Shopify, Etsy, Wayfair, and Target. UCP is endorsed by 20+ partners including Stripe, Visa, Mastercard, Adyen, American Express, Best Buy, Macy's, Home Depot, and Zalando (Google, Jan 2026).

Amazon Rufus

Amazon's AI shopping assistant Rufus serves 300 million users and drove an estimated $12 billion in incremental sales in 2025. Amazon's “Buy for Me” feature attempts to purchase products from third-party websites on behalf of consumers (Business Insider, Jan 2026).

Perplexity Shopping

Perplexity's PayPal-powered shopping feature enables instant purchasing within search results. With over 45 million monthly active users and a $20 billion valuation, Perplexity has become a significant AI commerce channel (Perplexity, 2025).

Beauty and Skincare

Beauty is the proving ground. Ulta Beauty went live on UCP as an early launch partner. Glossier ships through ACP on ChatGPT. SKIMS and Vuori are building for AI-first discovery with structured product catalogs and conversational checkout-ready feeds. Beauty wins because consumer queries map cleanly to structured product attributes (skin type, finish, ingredient, concern) and brand-direct purchase behavior is established.

Fashion and Apparel

Spanx, Vuori, Coach, Kate Spade, Revolve, Halara, Abt Electronics, and the URBN family (Anthropologie, Free People, Urban Outfitters) are live on Stripe's Agentic Commerce Suite. Etsy was the first major marketplace to integrate ACP and remains the largest catalog source for ChatGPT Shopping's handmade and vintage inventory. Apparel retailers see conversational queries like “black wool coat under $300, ships by Friday” that traditional search struggles to interpret.

Grocery and Local Commerce

Albertsons is building a proprietary AI shopping assistant aimed at reducing average shopping time from 46 minutes to 4 minutes. DoorDash expanded its ChatGPT app beyond groceries into every local commerce category. Grocery is harder than fashion because inventory turns hourly, but the conversion-rate advantage is large: AI agents can auto-substitute and resolve out-of-stock issues without breaking the cart.

Home Goods and Furniture

Wayfair, Ashley Furniture, and Home Depot are UCP launch partners. Best Buy joined as a Google AI Mode partner. Furniture is a high-consideration vertical: AI agents excel at multi-attribute filtering (dimensions, room fit, style, delivery window) that traditional faceted search makes painful.

Marketplaces and Mid-Market Retail

Walmart and Target were announced as upcoming ChatGPT Shopping partners. Walmart announced AI training for 2.1 million associates. Macy's endorsed UCP at launch. Mid-market retailers without enterprise PIM systems are the largest underserved segment. Many are auto-enrolled in ChatGPT Shopping and Google AI Mode via Shopify but most products don't show up because their catalogs aren't structured for AI agents to evaluate.

The protocol landscape is converging. UCP was explicitly designed to be compatible with AP2, A2A, and MCP. A retailer who supports UCP and ACP covers the major AI shopping channels today. Infrastructure platforms like Paz.ai handle multi-protocol compliance through a single integration, translating merchant catalogs into the format each protocol requires.

Myth: Customers buy inside ChatGPT.

Reality: OpenAI deprecated Instant Checkout in March 2026. The current ACP model is product discovery plus merchant redirect. AI agents recommend; shoppers buy on the retailer's site. Brands keep customer relationships, login data, and loyalty engagement. Vendor pages still describing in-chat purchases are stale.

Myth: A great PIM is enough.

Reality: Internal PIM workflows and external agentic readiness are different problems. Structured catalog quality at the protocol layer is what AI agents evaluate. A retailer with a sophisticated PIM and a poor protocol-compliant feed gets skipped. A retailer with a basic PIM and a clean ACP/UCP feed gets recommended.

Myth: It's only for large retailers.

Reality: Over 1 million Shopify merchants are auto-enrolled in ChatGPT Shopping and Google AI Mode. The question for mid-market retailers isn't whether to participate - they already are. It's whether their products are structured well enough to actually appear.

Myth: Compliance with one protocol is enough.

Reality: Ten protocols are live in 2026 with zero interoperability. Microsoft adopted UCP for Copilot. ChatGPT runs on ACP. A retailer wanting both major surfaces needs both. UCP and ACP are explicitly designed to be complementary, not competing.

Myth: AI shopping is hype, not revenue.

Reality: AI-referred shoppers converted 42% better than human shoppers in Adobe's Q1 2026 data, and AI traffic to U.S. retailers grew 393% YoY. Retailers with AI agent integrations saw 7x sales growth during Cyber Week 2025. The category is small in absolute terms but growing faster than any prior commerce shift.

Myth: Compliance is the moat.

Reality: Compliance is the floor. Once your catalog is ACP and UCP compliant, every competitor in your vertical can match the same checkbox. The durable advantage is ongoing optimization based on how AI engines actually rank your products: which queries you win, which you lose, and which catalog changes move the needle across ChatGPT, Google AI Mode, and Perplexity over time.

Approach 1: Platform-Native (Fastest)

Shopify merchants are already auto-enrolled in ChatGPT shopping and Google AI Mode through their existing Shopify feeds. This provides baseline visibility with zero additional work. The tradeoff: no optimization, no branded experience, and limited control over how products appear to AI agents.

Approach 2: Infrastructure Partner (Balanced)

Platforms like Paz.ai connect existing product catalogs to multiple AI shopping channels through a single integration. This approach adds catalog optimization (rewriting product descriptions for AI comprehension), multi-protocol compliance, and analytics — typically going live in 2-4 weeks.

Approach 3: Custom Build (Maximum Control)

Large retailers like Albertsons have built proprietary AI shopping assistants integrated with their loyalty programs, inventory systems, and fulfillment operations. Albertsons aims to reduce average shopping time from 46 minutes to 4 minutes. This approach requires significant engineering investment ($10M+) and 12-18 months of development, but delivers maximum differentiation.

Most mid-market retailers will start with Approach 1 or 2 and layer on custom capabilities over time. The critical first step is ensuring your product catalog is machine-readable — AI agents cannot recommend products they cannot understand.

Found rate

The percent of relevant queries where your products appear at all in AI-engine results. A 0% found rate on a high-intent query like “best running shoes for flat feet” means agents never see you. Found rate is the entry-level signal: nothing else matters until this is non-zero.

Position quality

When you do appear, are you a primary recommendation or a footnote? Position 1-3 in an AI-engine carousel drives material click-through; position 7+ is essentially invisible. Track position by query, by engine (ChatGPT, Google AI Mode, Perplexity), and over time.

Share of voice vs named competitors

For every query in your category, what percent of recommendations go to you versus each named competitor? This is the canonical GEO metric. Watch trend lines, not absolute numbers.

Catalog completeness score

What percent of your SKUs have all the structured attributes AI agents evaluate? Most attribute completeness rates vary widely across retailers and verticals on existing catalogs. Closing this gap is the single highest-leverage operational improvement most teams can make. Run an AI Readiness check to see your current score.

AI-referred conversion rate

Adobe Analytics measured 42% higher conversion among AI-referred shoppers in Q1 2026 compared to traditional traffic. Track this on your own site - if your AI-referred conversion rate isn't higher than your search-referred conversion rate, your product detail pages aren't handling the high-intent traffic well.

Citation rate in long-form AI answers

When ChatGPT, Claude, or Perplexity answers a buying-related question with multiple paragraphs of guidance, are you cited as a source? AEO citation rate is the harder cousin of GEO share of voice - it tracks credibility, not just product visibility.

The most consequential shift: in traditional ecommerce, brands compete for consumer attention through design, advertising, and search ranking. In agentic commerce, brands compete for AI agent selection through data quality, protocol compliance, and catalog completeness.

This is why AI visibility — the discipline of making products discoverable and recommendable by AI agents — is becoming as critical as SEO was for the search engine era.

Agentic commerce is live, growing, and already generating measurable revenue for retailers who have adopted it. With 50 million daily shopping queries on ChatGPT alone, 805% year-over-year AI traffic growth, and a projected $3-5 trillion global market by 2030, the question is not whether to prepare — it is how quickly you can act.

The retailers winning today share three traits: their product catalogs are machine-readable, they support multiple AI commerce protocols, and they optimize their data for AI agent comprehension rather than just human browsing.