Agentic commerce means AI agents doing part of the shopping for people: finding products, comparing them, checking reviews, stock and delivery times, and in some cases placing the order with the shopper's permission. The 2026 holiday season is widely described as the first real test of it.
Here's what's actually happening, what's hype and what shops should do now.
Where things stand in late 2026
| Development | Status |
|---|---|
| AI-driven traffic to retail sites | Growing fast; Adobe forecasts a large year-on-year jump for the 2026 holidays |
| ChatGPT shopping | Product recommendations and links; in-chat Instant Checkout launched in 2025 and was scaled back in March 2026, so most purchases finish on merchant sites |
| Shopify Agentic Storefronts | Lists merchants' products in AI assistants such as ChatGPT, Microsoft Copilot and Perplexity |
| Amazon "Buy for Me" | Lets Amazon's agent complete purchases on some other retailers' sites |
| Walmart, Google and others | Bringing shopping assistants into AI chat and AI Mode |
| Agent-completed checkout share | Still small; most AI influence is in discovery and shortlisting |
The honest picture: AI agents already shape what people consider buying. Fully agent-completed purchases are still a small slice. Both matter.
How an AI shopping agent decides
When someone asks "Find me a waterproof hiking jacket under £150 that arrives before Saturday," an agent typically:
- Breaks the request into criteria: waterproof rating, price, size, delivery date.
- Searches product catalogues, feeds, retailer sites and reviews.
- Filters out products that don't clearly meet every criterion.
- Compares the rest on price, reviews, delivery and returns.
- Presents a shortlist, or completes the purchase if allowed.
Notice step 3. If your product page doesn't clearly state the waterproof rating, delivery time or returns policy, the agent can't confirm it fits, and you're dropped. Missing information is the new invisibility.
What shops should do now
1. Make product data complete and precise
- Specific titles with brand, product type and key attributes.
- Every relevant attribute filled in: size, colour, material, dimensions, compatibility, ratings.
- Accurate price and stock, updated in real time.
- GTINs and brand where applicable.
- Several high-quality images.
Read product feed optimisation for Google Shopping and AI assistants.
2. Add Product structured data
Mark up price, availability, shipping, returns and reviews on product pages. Agents and search engines both read it. Our schema generator helps.
3. State delivery and returns clearly
Delivery cut-off times, costs, delivery estimates and returns terms should be on the product page, not hidden in a policy page.
4. Join the channels that list you in AI assistants
- On Shopify, check your Agentic Storefronts settings and the Shopify Catalog. See Shopify agentic storefronts.
- Keep your Google Merchant Center feed healthy for AI Mode and Shopping.
- Consider marketplaces where agents shop.
5. Earn reviews and mentions
Agents weigh reviews and reputation heavily. A steady flow of genuine reviews and mentions on trusted sites helps. Zelvorm Loop collects reviews after delivery. Read brand mentions for AI search.
6. Make your site agent-friendly
Accessible pages, clear forms and fast load times help agents that browse your site directly. Look at WebMCP and our agent-ready website checklist.
7. Keep checkout fast and trustworthy
Most AI-assisted journeys still end on your checkout. Express payment options, guest checkout and clear costs matter more than ever. See cart abandonment fixes.
How to measure it
- Track referrals from AI assistants in GA4 with a custom channel. Our guide to tracking AI traffic shows how.
- Watch Shopify's channel reports if you use Agentic Storefronts.
- Check which products AI assistants recommend for your key searches each month.
Risks and open questions
- Fraud and disputes: agent purchases raise new questions about authorisation. Use platforms with clear rules and fraud checks. For cash-on-delivery markets, see COD fraud and RTO.
- Margin pressure: agents compare prices ruthlessly. Compete on value, delivery and trust, not just price. Zelvorm Price Lab helps find the right price.
- Brand distance: if agents do the shopping, customers may never see your brand story. Build direct relationships through email and great post-purchase experiences.
The short version
Agentic commerce is real in discovery and comparison, and growing in checkout. AI agents drop products they can't verify, so complete product data, clean feeds, structured data and clear delivery and returns information are now essential. Join the channels that list you in AI assistants, earn reviews and keep checkout fast.
Want your shop ready for AI shoppers? Our e-commerce team can audit your feed, product pages and checkout. Running a shop day to day? Zelvorm Commerce brings orders, products and ads together in one place.




