ChatGPT's Instant Checkout and Perplexity's PayPal-powered in-chat purchasing mean AI agents can now find a product, compare it, and complete a purchase without the shopper ever landing on your website. For marketers, this means product and pricing data now has two audiences: humans browsing your site, and AI agents parsing your structured data to decide whether to recommend and transact with you at all. If your product pages aren't machine-readable in the right way, you're invisible to a growing slice of purchase intent, regardless of how good your page looks to a human.
What "Agentic Commerce" Actually Means
Agentic commerce is the general term for AI systems completing (not just recommending) a purchase on a user's behalf. ChatGPT's Instant Checkout, launched as part of OpenAI's push into shopping, lets users buy products directly inside a chat conversation. Perplexity's PayPal integration does something similar, enabling in-chat payment for products surfaced in its answers. A wave of retailers adopted these integrations through 2026, and the direction of travel is clear: more purchase journeys are starting and ending inside an AI chat interface, not a browser tab.
This is a meaningful shift from traditional SEO or even standard AEO. It's not enough to be findable and readable, your product data has to be structured cleanly enough that an agent can confidently extract price, availability, and trust signals to complete a transaction without human back-and-forth.
What AI Shopping Agents Need to "See" Clearly
Structured product data (schema.org markup)
This is the foundation. If your Product, Offer, and AggregateRating schema isn't implemented correctly, agents may not reliably pull your pricing or availability at all. At minimum, every purchasable product page should have:
Productschema with name, description, brand, and SKU/GTIN where applicableOfferschema with explicit price, currency, and availability status (InStock, OutOfStock, PreOrder)AggregateRatingandReviewschema reflecting real customer feedback- Valid, current
priceValidUntildates so agents don't surface stale pricing
Unambiguous pricing
Agents need a single, clear price signal. If your pricing display relies on JavaScript-rendered discounts, region-based dynamic pricing that isn't reflected in markup, or vague "starting at" language without a structured minimum price, that ambiguity works against you. Make the actual transactable price explicit in both the visible page and the schema.
Clear availability and fulfillment signals
Agents making a purchase decision on a user's behalf need confidence the item is actually available and will ship as promised. Keep inventory status accurate in real time, and make shipping/delivery windows explicit rather than buried in a separate policy page.
Review and trust signals in a parseable format
Star ratings and review counts embedded in structured data (not just rendered visually) give agents a trust signal they can compare across competing products. If a competitor has schema-embedded reviews and you don't, an agent comparing options may simply never surface you as an option.
Clean, canonical product URLs
Agentic checkout flows need to resolve to a stable, canonical URL for each product/offer. Faceted navigation URLs, session-based parameters, or inconsistent redirects can break an agent's ability to complete a transaction cleanly.
An Audit Checklist for Your Product and Pricing Pages
- Validate your Product and Offer schema with Google's Rich Results Test and Schema.org's validator, don't assume it's correct just because it was implemented once.
- Confirm price and availability in your schema match what's actually rendered and actually true in your inventory system, checked live, not from a stale export.
- Check that reviews/ratings are marked up in structured data, not just displayed as star icons.
- Test whether your product pages are crawlable and indexable at all, agentic commerce still depends on the underlying content being accessible, not blocked by robots.txt or requiring login.
- Simplify checkout-adjacent information (returns, shipping cost, delivery time) into clear, extractable statements rather than long policy prose.
- Where possible, look into whether your platform (Shopify, WooCommerce, custom stack) has direct integrations or feeds for ChatGPT Instant Checkout or Perplexity's commerce partners, since some of this is increasingly handled via product feed integrations rather than page scraping alone.
What's Still Uncertain
The specific technical requirements for Instant Checkout and Perplexity's PayPal flow are still evolving, and not every retailer or platform has documented integration paths as of this writing. Some of what's needed may come through structured product feeds (similar to Google Shopping feeds) rather than purely on-page schema, this is an area worth monitoring rather than treating as fully settled. If your business handles significant transaction volume, it's worth checking directly with OpenAI's and Perplexity's current merchant documentation rather than relying solely on general best practices, since these programs are still expanding their partner lists.
Why This Matters Even If You're Not on ChatGPT or Perplexity's Merchant List Yet
Even if your business hasn't set up a formal Instant Checkout or Perplexity commerce integration, AI shopping agents can still surface your products in comparison and recommendation contexts by reading your public pages and structured data, the "research and recommend" function works independently of the "complete the transaction" function. That means product page optimization for AI discoverability is worth doing now, ahead of any specific checkout integration, because it affects whether you're even part of the consideration set an agent presents to a shopper.
Think of it in two layers: the first layer is being accurately findable, comparable, and citable by an AI agent doing research on a user's behalf. The second layer is being technically capable of completing a transaction through a formal integration. Most businesses should prioritize the first layer immediately, since it requires no partnership agreement, no new technical integration, and directly overlaps with SEO and AEO best practices you likely already have reason to invest in.
Common Mistakes That Quietly Break Agent Discoverability
Pricing rendered only via JavaScript with no server-side fallback. Many crawlers and AI agents don't fully execute JavaScript the way a browser does. If your price only appears after a script runs, some agents may see a blank or "price unavailable" state.
Inconsistent pricing across schema and visible page. If a promotional banner shows one price but your Offer schema still reflects list price, agents may cite the wrong number, creating a mismatch the shopper discovers at checkout, which damages trust in your brand, not just the specific transaction.
Missing or expired `priceValidUntil` fields. Stale schema-declared pricing can get pulled into an AI answer well after a sale has ended, creating a bad experience and a credibility hit.
Reviews displayed as images or icon-only star ratings with no underlying markup. If the actual review count and rating aren't in the HTML or schema, agents effectively see no reviews at all, even if your product page looks trustworthy to a human.
Regional or membership-gated pricing with no default state. If price varies by login status or region and there's no clear default/public price available to an unauthenticated crawler, agents may simply skip the page.
FAQ
What is agentic commerce? Agentic commerce refers to AI systems (like ChatGPT or Perplexity) completing a purchase on a user's behalf within a chat interface, rather than just recommending a product for the user to buy elsewhere.
Do I need special integration to be found by AI shopping agents? At a baseline, clean structured data (Product, Offer, Review schema) helps agents parse your pages. For direct checkout integrations like ChatGPT's Instant Checkout, you may need a formal merchant/partner integration, which is still expanding as of 2026.
Does this replace traditional e-commerce SEO? No, traditional SEO and structured data best practices are the foundation agentic commerce optimization builds on. Sites with weak technical SEO were already at a disadvantage; agentic commerce raises the stakes.
How do I know if AI agents are already finding my products? Check referral traffic sources in your analytics for patterns tied to AI platforms, and manually test by asking ChatGPT or Perplexity to find and compare products in your category to see if you appear.
Is this only relevant for e-commerce, or does it matter for service-based businesses too? It's most directly relevant to businesses selling discrete, priced products or clearly packaged offers (this includes course/program pricing for edtech, which I cover in a follow-up post). Service businesses should still ensure pricing and offer information is structured clearly, since AI agents are increasingly used for research even when they can't complete the transaction directly.
Getting the fundamentals right, clean structured data, clear offers, real trust signals, is the same groundwork that makes both traditional SEO and this new agentic layer work. It's the kind of technical-plus-strategy audit I do for edtech and startup clients navigating exactly this shift.