What Is Agentic Commerce? A Complete Guide for Ecommerce Brands

McKinsey estimates that agentic commerce could orchestrate $3 to $5 trillion in global retail revenue by 2030. That's not a typo. AI agents that research, compare, and buy products on behalf of consumers are moving from concept to reality faster than most ecommerce teams expected. Google, OpenAI, Stripe, Visa, Mastercard, and PayPal have all shipped agentic commerce infrastructure in the past twelve months, and Shopify reports that orders from AI-powered searches grew 15x year-over-year through 2025.

If you run an ecommerce brand, this shift changes how your products get discovered, evaluated, and purchased. This guide breaks down what agentic commerce actually is, how the major protocols work, what the numbers look like, and how to make sure your store is ready.

What Is Agentic Commerce?

Agentic commerce is a model of digital buying where AI agents act on behalf of consumers or businesses to research, compare, and complete purchases, often without direct human involvement. IBM defines it as "an approach to buying and selling in which AI agents act on behalf of consumers or businesses to research, negotiate and complete purchases, often without direct human intervention."

The key word is autonomy. Traditional ecommerce requires you to search, browse, compare, and click "buy" yourself. Conversational commerce (chatbots, AI assistants) recommends products and answers questions, but you still make the final call. Agentic commerce goes further: the AI agent handles the entire workflow, from finding the right product to completing checkout.

Think of it this way. You tell an AI agent, "Find me a moisturizer for dry skin under $40 with free shipping." The agent searches across multiple retailers, reads reviews, compares ingredients, checks your past preferences, and places the order. You approve the final step, or in some cases, the agent handles that too.

Forrester draws a clear line between the two paradigms: conversational commerce involves "chatting with AI assistants that suggest products," while agentic commerce "delegates the actual buying process to AI agents." For a deeper look at how conversational and agentic approaches compare, see our conversational commerce guide.

Three Capabilities That Define an Agentic System

Not every AI chatbot qualifies as an agentic commerce system. According to IBM, three capabilities separate true agentic systems from standard automation:

• Autonomy: The agent acts without constant user input. It doesn't wait for you to click "next" at every step.
• Reasoning: The agent adapts to changing conditions. If a product goes out of stock or a price drops, it adjusts its plan.
• Interoperability: The agent connects across platforms, retailers, and payment systems through open APIs and protocols.

This last point, interoperability, is where the industry has made the most progress in 2026. Two competing-but-complementary open protocols now give AI agents a standardized way to talk to merchant systems.

How Agentic Commerce Works: From Intent to Purchase

The flow of an agentic transaction follows three stages, regardless of which protocol or AI platform powers it.

Stage 1: Intent Recognition

The consumer defines a goal, along with permissions and constraints. "Find me a camping tent under $150 that fits four people and ships by Friday" is a complete intent statement. The agent parses this into actionable parameters: product category, price ceiling, size, and delivery timeline.

Stage 2: Reasoning and Planning

The agent builds a multi-step plan. It queries product catalogs across retailers, compares specifications and pricing, reads reviews, checks inventory in real time, and applies any available coupons or loyalty rewards. If a top-ranked option goes out of stock mid-search, the agent re-ranks alternatives without starting over.

Stage 3: Autonomous Execution

The agent completes the transaction. Depending on the consumer's trust settings, this could mean presenting a final recommendation for approval or executing the purchase directly. Post-purchase, the agent can track shipping, handle returns, and schedule reorders.

Today, most agentic systems still include a human approval step before payment. A Riskified study from Q1 2026 found that 55% of consumers aren't comfortable letting AI agents make purchases without their sign-off. But 47% of US shoppers already use AI tools for at least one shopping task, according to Visa. The gap between discovery and autonomous buying is closing fast.

The Protocols Powering Agentic Commerce

For AI agents to buy from merchants, both sides need a shared language. Two open protocols have emerged as the foundation, backed by the biggest names in tech and payments. For a technical deep-dive on how these protocols compare, read our ACP vs UCP vs MCP breakdown.

Universal Commerce Protocol (UCP): Google + Shopify

Google and Shopify co-developed UCP and announced it at NRF in January 2026. It's an open-source standard built on REST and JSON-RPC that lets AI agents connect to any merchant's catalog, initiate checkout, and manage orders through a unified API.

UCP launched with three core capabilities: Checkout, Identity Linking, and Order Management. Its partner list reads like a who's-who of retail: Walmart, Target, Best Buy, Home Depot, Macy's, Etsy, Wayfair, Flipkart, and Zalando. Payment networks Visa, Mastercard, American Express, Stripe, and Adyen have all signed on.

The first live implementation powers buying inside Google AI Mode Search and Gemini. When you search for a product in Google's AI mode, the agent can now complete the purchase without leaving the search interface. For more on how UCP affects your store, see our UCP guide for ecommerce brands.

Agentic Commerce Protocol (ACP): OpenAI + Stripe

OpenAI and Stripe launched ACP in September 2025 as an open standard (Apache 2.0 license) that defines how AI agents and merchants coordinate checkout and share payment credentials securely. The merchant stays the merchant of record, keeping full control over products, pricing, and fulfillment.

ACP powers ChatGPT Instant Checkout, which went live with Etsy first and is expanding to over one million Shopify merchants, including Glossier, Vuori, Spanx, and SKIMS. Microsoft's Copilot Checkout also runs on ACP, covering Bing, MSN, and Edge.

Stripe introduced Shared Payment Tokens (SPTs) as a new payment primitive for agent transactions, giving consumers a way to pay through AI agents without exposing their full card details to every merchant.

The Payment Networks: Visa, Mastercard, and PayPal

Protocols handle the "what to buy" layer. Payment networks handle the "how to pay" layer.

Mastercard launched Agent Pay in April 2025, using Agentic Tokens built on its existing tokenization infrastructure. Agent Pay works through standard card payment fields, meaning merchants don't need new code to accept agentic payments. Mastercard also partnered with PayPal to integrate Agent Pay into PayPal's wallet, covering hundreds of millions of consumers.

Visa launched Intelligent Commerce with 100+ partners globally and unveiled Intelligent Commerce Connect, a cross-network platform that supports ACP, UCP, and other protocols. Visa predicts millions of consumers will use AI agents for purchases by the 2026 holiday season.

PayPal rolled out its "Agent Ready" solution in October 2025, making its existing merchant base accessible to AI agents. PayPal also powers checkout for Microsoft Copilot and Perplexity's "Buy with Pro" feature. Its multiyear partnership with Google Cloud connects PayPal payments to Google's conversational commerce agent.

Agentic Commerce by the Numbers

The market data tells a consistent story: early-stage today, massive by 2030.

• $5.71 billion: Global agentic commerce market in 2025, projected to reach $65.47 billion by 2033 at a 35.7% CAGR (Grand View Research)
• $190 to $385 billion: Estimated US ecommerce spending through agents by 2030 (Morgan Stanley)
• 15% to 25%: Share of ecommerce expected to flow through agentic channels by 2030 (Bain)
• 45%: Consumers who already use AI for part of the buying journey (IBM IBV 2026 study)
• 805%: Year-over-year increase in AI traffic to US retail sites on Black Friday 2025 (Adobe)
• 4.4x: Higher conversion rates from AI-generated product recommendations versus traditional search (McKinsey)

But here's the nuance. Less than 0.2% of ecommerce sessions currently come from ChatGPT referrals, and those referrals convert 86% worse than affiliate links. The infrastructure is being built, but the transaction volume is still small. Brands that prepare now will capture the growth curve. Brands that wait may find themselves invisible to AI agents entirely.

Microsoft reported that Copilot shopping journeys led to 53% more purchases within 30 minutes of interaction, showing that when the AI surfaces match high-intent shoppers, conversion rates spike. For context on how AI-assisted checkout compares to standard flows, see our data on why AI-assisted shoppers complete checkout at nearly 2x the rate.

How Ecommerce Brands Should Prepare for Agentic Commerce

You don't need to rebuild your store from scratch. But you do need to make your products machine-readable and your systems agent-friendly. Here's what matters most.

Get Your Product Data in Order

AI agents don't browse your site the way humans do. They read structured data. Product titles, descriptions, prices, dimensions, materials, availability, and shipping details all need to be clean, complete, and formatted in standard schemas (JSON-LD with Schema.org markup).

This isn't optional. Mirakl reports that 42% of customers abandon purchases due to insufficient product information, and brands lose an average of $15 million annually from poor data quality. When AI agents evaluate your catalog, missing or inconsistent data means your products get skipped. For detailed guidance on which fields matter most, check our post on what product data fields actually matter for AI shopping recommendations.

Start Thinking About Generative Engine Optimization (GEO)

SEO gets your products found by humans on Google. GEO gets your products found by AI agents on ChatGPT, Copilot, Gemini, and Perplexity. The two overlap (structured data helps both), but GEO also involves machine-readable policy pages, clear return and shipping terms, and accurate inventory feeds.

HBR recommends adopting llms.txt standards to make your site AI-accessible, and monitoring your "share of model," which tracks how AI agents represent your brand across different platforms. Pernod Ricard pioneered this approach for the spirits category.

Connect to the Protocol Ecosystem

If you're on Shopify, you're already close to both protocols. Shopify's Agentic Storefronts let merchants set up their data once and get surfaced across ChatGPT, Copilot, Google AI Mode, and Gemini. Shopify even launched an "Agentic plan" that lets brands on other platforms use its infrastructure for AI channels.

For brands on other platforms, Stripe's Agentic Commerce Suite supports WooCommerce, BigCommerce, Squarespace, and commercetools. Feedonomics launched Agentic Catalog Exports (ACE) to syndicate catalog data directly to OpenAI and Google Gemini. Our guide on agentic storefronts covers the setup process for each major platform.

Prepare for Agentic Payments

Make sure your payment stack supports tokenized transactions. Mastercard's Agent Pay, Visa's Intelligent Commerce, and Stripe's Shared Payment Tokens all work through existing card rails, so most merchants won't need a rip-and-replace. But check with your payment processor to confirm they support agentic payment flows.

How Alhena AI Helps Brands Get Ready for Agentic Commerce

While industry protocols define how external AI agents (ChatGPT, Copilot, Gemini) talk to your store, you also need your own AI agent working for you on your own channels. That's where Alhena AI's Shopping Assistant fits in.

Alhena gives ecommerce brands an AI agent that already handles the core agentic commerce workflow: product discovery, guided recommendations, cart population, and checkout assistance. It's purpose-built for ecommerce sales, not retrofitted from a support ticket tool.

Your Brand's Own AI Agent

Alhena's Product Expert Agent understands natural language queries like "I need a moisturizer for sensitive skin under $35" and returns personalized recommendations grounded in your actual product catalog. No hallucinated products, no made-up specs. The agent populates carts, applies discounts, and walks shoppers through checkout.

Tatcha saw a 3x conversion rate and 38% AOV uplift after deploying Alhena, with 11.4% of total site revenue attributed to AI conversations. Victoria Beckham Beauty reported a 20% AOV increase. These results come from an AI agent that sells, not just one that deflects tickets. See the full Tatcha case study for details.

Omnichannel by Default

Agentic commerce won't be limited to web chat. Consumers will interact with AI agents through Instagram DMs, WhatsApp, email, and voice. Alhena's Social Commerce and Voice AI products cover all of these channels with a single unified agent, so your brand voice and product knowledge stay consistent regardless of where the conversation happens.

Built-In Revenue Attribution

One of the biggest challenges in agentic commerce is measuring what the AI actually drove. Alhena includes revenue attribution analytics that track which conversations led to purchases, what products the AI recommended, and how AI-assisted shoppers compare to unassisted ones. You don't need to guess whether the AI is working. The ROI calculator shows projected impact before you even start.

Works With Your Existing Stack

Alhena integrates with Shopify, WooCommerce, Magento, and Salesforce Commerce Cloud for product data. For support workflows, it connects to Zendesk, Freshdesk, Gorgias, Intercom, and more. Deployment takes under 48 hours with no dev resources required. Brands like Puffy achieved 63% automated inquiry resolution and 90% CSAT while Crocus hit an 86% deflection rate with 84% CSAT.

Key Takeaways

• Agentic commerce means AI agents that autonomously research, compare, and buy products on behalf of consumers. It goes beyond chatbots and conversational commerce.
• Two open protocols (Google's UCP and OpenAI/Stripe's ACP) now provide the infrastructure for AI agents to transact with merchants at scale.
• Payment networks (Visa, Mastercard, PayPal) have built agentic payment layers that work through existing card rails, lowering the adoption barrier for merchants.
• The market is early but growing fast: 45% of consumers already use AI in their buying journey, and McKinsey projects $3 to $5 trillion in global agentic commerce revenue by 2030.
• Product data quality is the single most important preparation step. AI agents skip products with incomplete or inconsistent data.
• Brands need their own AI agent on their own channels. Alhena AI gives ecommerce brands a sales-focused agent that handles discovery, recommendations, cart building, and checkout across web, social, email, and voice.

Ready to give your store its own AI shopping agent? Book a demo with Alhena AI or start free with 25 conversations.

Alhena AI

Schedule a Demo

Frequently Asked Questions