Agentic commerce is shopping in which an AI agent acts for a customer: it searches, compares, asks follow-up questions and, with permission, completes the purchase.
For twenty-five years, online shopping has followed the same path: search, click, browse, add to cart, check out. Agentic commerce changes the path. A shopper tells an AI assistant, “Find me waterproof hiking boots under $150 that ship before Friday,” and the agent does the searching, filtering and comparing. In some cases it completes the purchase without the shopper ever visiting the store’s website.
This guide explains what agentic commerce is, what has actually launched, what it means for retailers and brands in the US, UK and Canada, and how to prepare your store without chasing hype.
What Is Agentic Commerce in Simple Terms?
An AI agent is software that can take actions towards a goal, not just answer questions. In commerce, that means an agent can:
- Understand intent: budget, preferences, sizes and deadlines.
- Discover products across many stores using feeds, websites and APIs.
- Compare specifications, reviews, prices, delivery times and return policies.
- Transact using stored payment credentials, with the user’s approval and within set limits.
- Follow up on delivery, returns or reorders.
In agentic commerce, your first customer is often a machine reading your data. If the machine cannot understand your product, the human never hears about it.
What Has Actually Launched?
| Development | What it is | Why it matters |
|---|---|---|
| ChatGPT Instant Checkout (2025) | Buying directly inside ChatGPT, starting with US Etsy sellers and expanding to Shopify merchants | Purchases can happen without visiting the merchant’s site |
| Agentic Commerce Protocol (ACP) | An open standard from OpenAI and Stripe for connecting merchants to AI agents | Merchants keep their payment processor and remain the merchant of record |
| Google Agent Payments Protocol (AP2) | An open protocol for secure, verifiable payments initiated by agents | Introduces signed “mandates” proving what the user authorised |
| Visa Intelligent Commerce and Mastercard Agent Pay | Card network programmes for agent-initiated payments using tokenised credentials | Builds trust and fraud controls into agent payments |
| AI shopping features in search | AI shopping in Google, Perplexity and others, drawing on product feeds and reviews | Product data quality decides visibility |
This space is moving quickly. Availability by country, platform and merchant category changes frequently, so check current eligibility with your ecommerce platform and payment provider before planning around a specific feature.
How Does an AI Agent Decide What to Buy?
Agents rely on information they can parse quickly and trust:
- Structured product data: titles, GTINs, brand, attributes (size, colour, material), price, availability and shipping.
- Reviews and ratings, both quantity and content.
- Policies: returns, warranties and delivery times.
- Merchant reputation and consistency across sources.
- Fit with the user’s constraints: price ceiling, delivery date and preferences.
Persuasive copy and beautiful photography still matter for humans, but agents reward completeness, accuracy and clarity.
Agent readiness is mostly data hygiene. The stores that win are the ones whose catalogue would already pass a strict Google Merchant Center audit.
How Do You Prepare Your Store for AI Shopping Agents?
1. Make product data complete and consistent
- Use precise titles: brand + product type + key attribute + variant.
- Fill every relevant attribute: material, dimensions, compatibility, care instructions, age range.
- Include GTINs or MPNs where they exist.
- Use the same data on your site, feeds and marketplaces.
2. Add Product and Offer structured data
Mark up price, currency, availability, shipping details, return policy and reviews using schema.org. Our Schema Generator creates valid JSON-LD.
3. Keep inventory and prices real-time
An agent that buys an item you cannot ship creates a refund, a support ticket and lost trust. Sync stock and prices frequently across feeds and channels.
4. Publish machine-readable policies
Write returns, shipping and warranty policies in clear, specific language: “Free returns within 30 days in the US, UK and Canada”, not “Hassle-free returns”.
5. Connect to the right platforms
Ask your ecommerce platform (Shopify, WooCommerce, BigCommerce and others) and payment provider (such as Stripe or PayPal) which agent checkout programmes they support, and opt in where it suits your margins and brand.
6. Answer questions on your product pages
Short FAQ sections such as “Is this machine washable?” or “Does it fit a 16-inch laptop?” feed both human shoppers and AI agents. See our guide to generative engine optimisation.
Run an “agent test”: ask ChatGPT, Gemini and Perplexity to recommend a product in your category with specific constraints. Note which stores they mention, what data they quote and what they get wrong about you. Then fix those data gaps first.
What Are the Benefits of Agentic Commerce for Merchants?
- New discovery channel beyond search ads and social media.
- Higher purchase intent: agents act on specific needs.
- Lower friction for repeat and replenishment purchases.
- Fair competition for specialists: well-described niche products can beat bigger brands on fit.
What Are the Risks?
- Less brand experience: buyers may never see your website.
- Customer data access: clarify what data you receive for orders and marketing consent.
- Platform fees and dependency on large AI platforms.
- Fraud and disputes: who is liable when an agent buys the wrong item?
- Price comparison pressure on commodity products.
Agentic Commerce vs Traditional Ecommerce: A Comparison
| Aspect | Traditional ecommerce | Agentic commerce |
|---|---|---|
| Discovery | Search ads, SEO, social, email | Agent queries across feeds, sites and reviews |
| Decision maker | Human browsing pages | Agent shortlists; human approves |
| Key asset | Website design and copy | Structured data, policies, reputation |
| Checkout | On the merchant’s site | Inside the AI interface, via protocols |
| Loyalty lever | Brand experience and email | Reliability, accurate data, post-purchase service |
Use Cases Across Industries
- Replenishment: groceries, pet food, office supplies and skincare that customers reorder regularly.
- Gift shopping: “Find a gift for a 10-year-old who loves science, under £40, delivered by Saturday.”
- Travel and events: agents booking hotels, restaurant tables or tickets within a budget.
- B2B procurement: agents reordering parts within approved suppliers and limits.
- Specialist retail: technical products where precise specifications decide the sale.
A small UK outdoor-gear retailer finds that AI assistants never mention its best waterproof jacket. Its product pages say “ultimate protection” but omit the waterproof rating, weight and seam taping. After the team adds those specifications, Product schema and a short FAQ, assistants can match the jacket to queries with exact requirements. The lesson for every store: vague marketing copy is invisible to agents.
Future Trends in Agentic Commerce
- Standardised protocols becoming part of every major commerce platform.
- Agent-to-agent negotiation, with buyer agents talking to merchant agents about stock, delivery and bundles.
- Verified merchant identity and trust signals designed for machines.
- Regulation clarifying consumer rights and liability for agent purchases in the US, UK, Canada and EU.
- Global adoption: fast-growing digital economies such as Saudi Arabia, where Vision 2030 is accelerating digital payments and ecommerce, are natural early markets for AI-assisted shopping.
Agentic Commerce Myths vs Facts
| Myth | Fact |
|---|---|
| “Agents will only buy the cheapest option.” | Agents follow user constraints, including quality, delivery date, brand preference and reviews, not just price. |
| “My website no longer matters.” | Your site remains the source of truth for product data, policies, content and repeat customers. |
| “This is only for big retailers.” | Small specialists with excellent product data can win specific, constraint-heavy queries. |
| “Agents bypass payment security.” | Emerging protocols use tokenised credentials, user authorisation and verifiable mandates designed to reduce fraud. |
| “I need to rebuild my store.” | Most readiness work is improving data, markup and policies on your existing platform. |
The practical message for most merchants: improve what you already have. Every fix to product data also lifts Google Shopping, marketplace listings and on-site search conversion today.
Your 60-Day Agent-Readiness Plan
| Weeks | Focus | Actions |
|---|---|---|
| 1–2 | Baseline | Run agent tests in major AI assistants, audit Merchant Center diagnostics, list missing product attributes |
| 3–4 | Data | Fix titles, attributes, GTINs and images for top-selling products first; standardise variants |
| 5–6 | Markup and policies | Add Product, Offer, shipping and return policy structured data; rewrite policies in specific language |
| 7–8 | Channels | Check platform and payment provider support for agent checkout; enable where margins allow; monitor orders and returns |
Common Agentic Commerce Mistakes
- Treating it as a marketing campaign instead of a data quality programme.
- Different prices across channels, which confuses agents and erodes trust.
- Hiding key facts in images, such as size charts uploaded only as pictures.
- Vague policies that agents cannot interpret reliably.
- Ignoring reviews: agents weigh ratings and review content heavily.
- Opting into every channel without checking fees, data sharing and returns handling.
Key Terms Explained
- AI agent: software that can plan and take actions towards a goal for a user.
- Merchant of record: the business legally responsible for the sale, payment and customer service.
- Product feed: a structured file or API listing products, prices and availability for channels.
- Tokenised credentials: payment details replaced by secure tokens so card numbers are not shared.
- Mandate (AP2): a signed record of what a user authorised an agent to buy.
What Can You Do This Week?
- Ask three AI assistants to recommend a product in your category with a budget and delivery date, and record whether your store appears.
- Export your product feed and sort by missing attributes; fix your top 20 sellers first.
- Rewrite your returns and shipping policies with exact days, costs and countries.
- Check your Product structured data with Google’s Rich Results Test.
- Ask your platform and payment provider which agent checkout programmes you are eligible for.
Small, specific steps taken this week create momentum, and they give you real data to plan the next phase with confidence.
Conclusion: Make Your Store Readable by Machines
Agentic commerce is the next shift in how people buy online. Shoppers increasingly delegate the search, comparison and even checkout to AI. Stores that win will have complete product data, accurate real-time stock and pricing, clear policies, strong reviews and connections to the right protocols and platforms. Start with data hygiene today. It improves your existing SEO and shopping performance now and prepares you for AI agents tomorrow.
Which Sources Back This Guide?
We checked the rules, rates and figures in this guide against the official and primary sources below. Fees, thresholds and platform features change, so confirm current details at the source before you act.
- Agentic Commerce Protocol: the open standard from OpenAI and Stripe.
- Agent Payments Protocol (AP2): Google's open protocol for agent payments.
- Google Search Central: Product structured data: markup that makes products machine-readable.
- Schema.org: Offer: price, availability and shipping properties.
Related guides on this site
How can we help with this?
Is your store ready for AI shopping agents?
We audit product data, structured markup, feeds and checkout integrations, then make your catalogue clear to both shoppers and AI agents.
Want this handled for you?
We build and optimise sites like this every week. Book a free 30-minute call and we will tell you exactly what we would change on yours.
Get a free site review
