Agentic commerce is shopping where an AI agent acts for the customer: it searches, compares products, checks price, stock, shipping and returns, and sometimes starts the purchase. To be chosen, your store needs to be readable by machines: crawler access, accurate Product and Offer structured data, clear policies, prices in the HTML and a good product feed.
This guide explains what the term means, what the forecasts for holiday 2026 actually say, a practical checklist you can work through before peak season, how to measure AI referrals in Google Analytics 4 and the risks to keep in mind. For the browser-side angle, our news piece on WebMCP and AI shopping agents covers the tools some platforms are exposing to agents.
Key takeaways
- Agentic commerce means AI agents finding, comparing and sometimes buying products for shoppers, using your product data instead of browsing like a person.
- Salesforce predicts 20% of holiday 2026 ecommerce traffic will come from AI chat agents, and Adobe forecasts AI traffic to US retail sites up 130% year over year.
- Agents need crawler access, Product and Offer schema, shipping and return policies, prices and stock in the HTML, and a clean product feed.
- Track AI referrals in GA4 with a custom channel for sources such as chatgpt.com and perplexity.ai, and keep checkout and policies under your control.
What does agentic commerce mean?
In ordinary ecommerce, a person visits your site, browses and checks out. In agentic commerce, part or all of that journey is done by software. A shopper might ask an assistant for "a waterproof hiking jacket under $200 that arrives before Christmas", and the agent reads product data from many stores, filters by price, size, delivery date and return terms, and presents a shortlist or starts checkout.
Salesforce describes the agents hitting store sites as a mix of "consumer-facing bots handling live queries", "autonomous agents executing backend tasks" and "competitor scrapers fueling algorithmic price-matching". It also notes that autonomous agents "rely on programmatic, real-time feed integrations" rather than browsing like humans.
Agentic commerce examples
- Product discovery in AI assistants. ChatGPT, Perplexity and similar tools recommend products and link to stores.
- Retailer shopping agents. A store's own assistant that answers questions and builds a basket.
- In-chat purchases through apps. Retailers building apps inside AI assistants so purchases happen through their own systems.
- Protocol-based checkout. Agents starting a purchase through a standard such as the Agentic Commerce Protocol.
What is the Agentic Commerce Protocol?
The Agentic Commerce Protocol (ACP) is described by its maintainers as "an open standard for programmatic commerce flows between buyers, AI agents, and businesses." It was developed by Stripe and OpenAI, is open source under Apache 2.0, and covers making checkout agent-ready and passing payment credentials without exposing the underlying details. OpenAI's ChatGPT is listed as the first AI platform to implement it and Stripe as the first compatible payment provider.
What's happening this holiday season
Here is what the main forecasts say. They are forecasts, not measured results.
| Source | Forecast | Notes |
|---|---|---|
| Salesforce (July 20, 2026) | "20% of all 2026 holiday ecommerce traffic will originate from AI chat agents" | Includes scrapers and backend bots, not only shoppers |
| Salesforce (July 20, 2026) | 1 in 3 ecommerce sites will have a site-specific shopper agent live by Cyber Week 2026 | Retailers' own agents |
| Adobe (September 28, 2026) | AI traffic to US retail sites up 130% year over year, Nov 1 to Dec 31 | Measured by shoppers clicking a link; Thanksgiving forecast up 159% |
| Adobe (September 28, 2026) | US online holiday spend of $275.1 billion, up 6.7% | Overall ecommerce, not AI-specific |
The ChatGPT shopping agent story so far
The path hasn't been straight. OpenAI launched Instant Checkout in ChatGPT in 2025. According to Retail Gazette in March 2026, OpenAI then scaled it back, saying the initial version "did not offer the level of flexibility that we aspire to provide" and that it would allow "merchants to use their own checkout experiences while we focus our efforts on product discovery." Merchants can share product feeds with OpenAI, and those wanting deeper integration can build apps inside ChatGPT.
The lesson for store owners: discovery is where most of the action is right now. Being findable and accurately described matters more than wiring up a new checkout.
Holiday 2026 agentic commerce checklist
Work through these in order. Our AI shopping agent checker tests a product page against most of them in one pass: crawler access in robots.txt, Product schema with offers, shipping and return policy schema, price in the HTML and llms.txt.
1. Let the right crawlers in (robots.txt)
AI assistants use separate crawlers for search, training and user-triggered visits. OpenAI says "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers," while GPTBot controls training use, and the settings are independent, so you can allow OAI-SearchBot and block GPTBot. Perplexity says PerplexityBot surfaces and links sites in its results and recommends allowing it. Check that your robots.txt, CDN and firewall aren't blocking these bots by accident.
2. Add Product and Offer structured data
Google's merchant listing documentation requires Product name, image and offers, with an Offer (not AggregateOffer) carrying price and priceCurrency. It recommends availability, brand, GTIN, sku, itemCondition and more. Agents filtering by price and stock rely on exactly these fields. Validate templates with a schema markup validator after any theme or app change.
3. Publish shipping and return policies in structured form
Google recommends a site-wide return policy and shipping policy under your Organization markup instead of repeating them on every product. Shipping policy markup uses ShippingService via hasShippingService on the page that describes your shipping policy; return policies use MerchantReturnPolicy, which requires applicableCountry and returnPolicyCategory. Since November 2025, Google also lets online merchants set shipping and returns in Search Console, and those settings take precedence over on-site structured data.
4. Put price and availability in the HTML
If price, stock or variant details only appear after JavaScript runs, some agents may not see them. Make sure the server-rendered page shows the price, currency, availability and key attributes such as size and color, and that they match your schema. Google Merchant Center asks you to match price and availability across your landing page, structured data and checkout.
5. Keep a clean product feed
Google Merchant Center's product data specification requires id, title, description, link, image_link, availability and price, plus brand for most new products. OpenAI publishes a product feed spec for ChatGPT with nine required fields, including seller_name, and tells merchants to update price and availability when they change. A feed that lags your site will show agents stale prices.
6. Write clear policy pages
Agents compare delivery cut-offs, return windows and who pays return shipping. Keep shipping, returns, warranty and contact pages plain, specific and dated, with holiday cut-off dates stated in text, not only in a banner image.
7. Consider llms.txt, but keep expectations low
llms.txt is a proposal from September 2024 for a Markdown file that points AI tools to your most useful content. It is an emerging convention, not a standard, and Google says you don't need new machine-readable or AI text files to appear in its AI features. Our llms.txt explainer covers when it is worth adding.
How to measure AI referrals in GA4
GA4 doesn't separate AI assistants by default; their visits usually land in Referral or Direct. Google's own help article on custom channel groups uses AI assistants as its example.
- In GA4 Admin, under Data display, open Channel groups and create a new group (copying the default).
- Add a channel called "AI assistants" with a condition on Source that matches regex, for example:
chatgpt|perplexity|copilot|gemini|claude. - Reorder so AI assistants sit above Referral. Google notes traffic goes to "the first channel whose definition it matches."
- Save, then view Traffic acquisition with your new channel group and compare revenue and conversion rate with other channels.
For a quick look without changing settings, open Reports → Acquisition → Traffic acquisition, switch to Session source, and filter for chatgpt.com, perplexity.ai and similar. Some AI visits arrive without a referrer and show as Direct, which is what happened in the October 2026 ChatGPT referral drop. To see whether assistants mention your brand at all, try our AI visibility checker.
Risks to plan for
- Stale or wrong data. An agent quoting yesterday's price or an out-of-stock size erodes trust and can create pricing disputes. Feed and schema accuracy is now a customer service issue.
- Scrapers mixed in with shoppers. Salesforce's own forecast counts competitor scrapers in its agent traffic. Opening up to AI search bots doesn't mean allowing unlimited scraping; use your CDN's bot controls.
- Fast-changing platforms. OpenAI's checkout change in 2026 shows features can be launched and pulled within months. Don't rebuild your checkout around one assistant.
- Payments and fraud. If you enable agent checkout, use your payment provider's supported tools and keep your normal fraud checks.
- Prompt injection and manipulated content. Agents read reviews and page text. Moderate user content and don't hide text meant only for bots.
- Over-investing in unproven conventions. Fix crawling, schema, feeds and policies before experimenting with llms.txt or new protocols.
If you'd like help connecting product data, feeds and order systems, our ecommerce automation hub shows where to start.
Frequently asked questions
What is the meaning of agentic commerce?
Agentic commerce is ecommerce where AI agents act for shoppers: searching, comparing and sometimes starting a purchase. Instead of a person browsing your pages, software reads your product data, policies and prices and decides what to recommend.
What are some agentic commerce examples?
Examples include AI assistants such as ChatGPT and Perplexity recommending products and linking to stores, retailers' own shopping agents on their sites, retailer apps inside AI assistants and protocol-based checkout using standards such as the Agentic Commerce Protocol.
What is the Agentic Commerce Protocol?
It is an open standard, developed by Stripe and OpenAI and released under Apache 2.0, for programmatic commerce between buyers, AI agents and businesses. It covers agent-ready checkout and passing payment credentials securely. ChatGPT is listed as the first AI platform to implement it.
Can the ChatGPT shopping agent buy from my store?
ChatGPT mainly helps shoppers discover products and links to stores. OpenAI scaled back its Instant Checkout in March 2026 in favor of merchants' own checkouts and apps inside ChatGPT. Allowing OAI-SearchBot and keeping product data accurate are the practical first steps.
Do I need to block or allow AI bots in robots.txt?
It depends on your goals. To appear in ChatGPT search answers, OpenAI says you need to allow OAI-SearchBot; you can still block GPTBot from training use. Perplexity recommends allowing PerplexityBot to appear in its results.
How do I see AI shopping traffic in Google Analytics?
Create a custom channel group in GA4 with a channel whose Source matches a regex such as chatgpt|perplexity|copilot|gemini|claude, and place it above Referral. Some AI visits arrive without a referrer and show as Direct, so treat the numbers as a minimum.
Sources
- Salesforce: 2026 Holiday Predictions (July 20, 2026)
- Adobe: U.S. holiday shopping season to hit record $275.1 billion online (September 28, 2026)
- Retail Gazette: OpenAI pivots ChatGPT shopping strategy after Instant Checkout underwhelms (March 2026)
- Agentic Commerce Protocol
- OpenAI: Overview of OpenAI crawlers
- Perplexity: Perplexity crawlers
- Google Search Central: Merchant listing structured data
- Google Search Central: Merchant shipping policy structured data
- Google Search Central Blog: More ways to share your shipping and returns policies (November 2025)
- Google Merchant Center: Product data specification
- OpenAI: Product feed spec
- llmstxt.org: The llms.txt proposal
- Google Search Central: AI features and your website
- Google Analytics Help: Custom channel groups
More guides for your industry
Free checklist (PDF)
Free e-commerce brands website checklist
30 practical checks for e-commerce brands: website speed and mobile, Google Business Profile and reviews, email security, booking and enquiry handling, and AI search visibility. Each one says why it matters, and many link to a free tool that checks it.
Written by the MakeItFlow team
We build the website audit these guides are based on. Every fact is checked against the primary sources listed above . How we research and check
Check your own site in 30 seconds
Run a free audit covering speed, SEO, DNS, email authentication and security headers — and get a prioritised fix list.