TL;DR
- $15 trillion in B2B spending will be intermediated by AI agents by 2028 (Gartner), and consumer shopping is following the same trajectory
- 64% of consumers plan to use AI chatbots for shopping in 2026, with nearly 1 in 4 making it their default method (PartnerCentric)
- ChatGPT, Google AI Mode, Perplexity, and OpenAI Operator are already enabling AI agents to discover, compare, and purchase products autonomously
- Products with complete Schema.org markup are 6.4x more likely to be selected by AI agents for recommendations (LLMRecommend.com, Q1 2026)
- Shopify launched Storefront MCP and Universal Commerce Protocol (UCP) to let AI agents search products, create carts, and initiate checkout
- Brands that become "agent-ready" now will capture disproportionate share as autonomous purchasing scales
What Is Agentic Commerce?
Agentic commerce is the shift from humans browsing and buying to AI agents autonomously discovering, comparing, and purchasing products on behalf of consumers.
This isn't a chatbot answering "Where's my order?" It's an AI agent that receives a goal ("Find me a sustainable sneaker brand under $150 with good arch support"), independently researches options across dozens of stores, evaluates fit based on your preferences and purchase history, and either recommends a shortlist or completes the purchase entirely.
The key distinction: traditional ecommerce requires the consumer to do the work (search, browse, compare, decide). Agentic commerce delegates that work to an AI agent that acts on the consumer's behalf.
The experience feels conversational and intuitive. A customer tells their AI agent what they need, and the agent handles everything from product discovery through transaction completion.
The AI Shopping Agents Reshaping Ecommerce
Four major platforms are already enabling agentic commerce at scale:
ChatGPT Shopping
Launched April 2025, ChatGPT Shopping delivers conversational product discovery across Free, Plus, and Pro tiers. Key characteristics:
- Organic, unsponsored results ranked purely on relevance (you cannot buy placement)
- Personalized recommendations with images, pricing, summarized reviews, and direct buy links
- Later added "Instant Checkout" for in-app purchasing
- 700 million weekly active users with access to shopping features
When someone asks ChatGPT "best moisturizer for dry skin under $40," your product either shows up in that conversation or it doesn't. There's no ad budget that fixes invisibility here.
Google AI Mode + "Buy For Me"
Announced at Google I/O in May 2025, Google AI Mode combines Gemini with the Shopping Graph containing 50 billion online products:
- Conversational product discovery with rich product cards
- AI virtual try-on (upload a photo, see clothing on your body)
- "Buy For Me": Gemini takes your payment info and completes the purchase autonomously on the retailer's site
- Triggers on price-tracking alerts when products hit user-defined thresholds
This is fully autonomous purchasing. The consumer doesn't visit your site, doesn't see your checkout flow, doesn't interact with your brand at all. The AI agent does everything.
Perplexity Buy with Pro
Launched November 2024, Perplexity's shopping integration provides:
- One-click checkout for Pro subscribers
- AI-generated product recommendations positioned as "unbiased"
- Free shipping on Buy with Pro orders
- Merchant Program for retailers to share product data directly
- Shopify and PayPal integrations
OpenAI Operator
Launched January 2025 as a Pro-only research preview:
- Vision-based agent that navigates websites like a human would
- Completes multi-step tasks: browsing, comparing, ordering, booking
- No API or integration required from the retailer's side
- Works on any website with a visible checkout flow
From Search-to-Buy to Agent-to-Buy
The traditional ecommerce funnel looks like this:
- Consumer searches a keyword
- Clicks through results
- Browses product pages
- Compares across tabs
- Decides and purchases
AI agents collapse this entire funnel into a single conversational interaction. The consumer states what they want, and the agent handles steps 1 through 5.
The data supports this shift:
- 64% of consumers plan to use AI chatbots for shopping in 2026 (PartnerCentric)
- Shopping-related GenAI use grew 35% from February to November 2025 (BCG)
- 30-45% of US consumers currently use GenAI for product research and comparison (Bain Consumer Lab)
- 17% of online shoppers will begin shopping journeys with AI platforms like ChatGPT or Perplexity (Bain)
- 30% plan to start with AI-enabled Google search (Bain)
Gen Z and Millennials are leading adoption. For these demographics, asking an AI agent to find the right product is becoming as natural as typing into a search bar.
The implication for brands: you no longer compete for clicks. You compete for agent recommendations. And the signals that win agent recommendations are fundamentally different from the signals that win Google rankings.
How AI Agents Discover and Choose Products
AI shopping agents don't browse your site the way humans do. They parse machine-readable data, evaluate trust signals, and make selection decisions based on structured information.
The key signals AI agents use to discover and recommend products:
1. Schema.org Product Markup
Products with complete Schema.org markup (Product, Offer, AggregateRating, Review, Brand) are 6.4x more likely to be selected in agent purchase decisions. This isn't a minor advantage. It's the difference between being visible and being invisible.
2. Reviews and Aggregate Ratings
AI agents treat review volume and sentiment as primary trust signals. High review counts with positive sentiment push products to the top of agent recommendations.
3. Competitive Pricing Data
Clearly structured pricing that agents can parse and compare. If your price isn't in a machine-readable format, agents can't include you in comparisons.
4. Product Feed Quality
Complete attributes, accurate inventory status, and up-to-date pricing in your product feeds. Incomplete feeds mean incomplete representation in agent results.
5. GTINs and Universal Identifiers
UPCs and GTINs help agents match and verify products across sources. Without them, agents can't confirm your product is the same one mentioned in reviews or comparisons.
6. Brand Trust Signals
Merchant credibility indicators, domain authority, and third-party validation. Agents prefer recommending from sources they can verify.
The critical insight: machine-readability, not visual design or traditional SEO alone, determines whether products surface in AI-driven commerce. Your beautiful product photography doesn't matter if your structured data is incomplete.
Shopify's Bet on Agent Commerce
Shopify isn't watching from the sidelines. Their Summer 2025 Editions release made their position clear:
Storefront MCP
Shopify's Storefront MCP enables AI agents to:
- Search products across your catalog
- Answer brand-specific questions
- Create carts
- Initiate checkout
It uses Model Context Protocol servers as intermediaries, connecting to live data including products, pricing, inventory, shipping, and store policies. As Shopify describes it: "Structured tools that read your store's real commercial state, so the model is not inventing."
Developer documentation is available at shopify.dev/docs/apps/build/storefront-mcp.
Universal Commerce Protocol (UCP)
Launched January 2026, UCP is an open standard co-developed with Google and supported by Target, Wayfair, and Etsy. It provides three layers:
- Universal primitives - standard product, cart, and order representations
- Standardized operations - common actions like discover, add-to-cart, checkout
- Custom extensions - brand-specific capabilities on top of the standard
UCP enables AI agents to conduct complete shopping transactions across any participating retailer using the same protocol.
Knowledge Base App
A merchant-facing tool that lets you customize the FAQs and information AI agents use when answering questions about your brand. This directly optimizes your store's visibility and accuracy in AI shopping conversations.
How to Prepare Your Store for AI Agents
An actionable checklist for making your store agent-ready:
1. Implement Complete Schema.org Product Markup
Every product needs Product, Offer, AggregateRating, Review, and Brand schema. Complete markup = 6.4x selection advantage.
2. Create or Connect to an MCP Server
Either build your own or use a pre-built solution like the Lexsis MCP Server to expose your store's capabilities to AI agents.
3. Optimize Product Feeds
Complete attributes, GTINs, competitive pricing, accurate inventory. Every missing field is a missed opportunity for agent discovery.
4. Build AI-Readable Content
Structured FAQs, clear product descriptions without marketing fluff, machine-parseable policies. Write for agents as much as humans.
5. Use Shopify's Knowledge Base App
Control what AI agents say about your brand. Don't leave your narrative to whatever the agent can scrape.
6. Maintain High Review Volume and Quality
Reviews are trust signals for agents. Actively collect and respond to reviews across platforms.
7. Ensure API Accessibility
Your product data should be accessible through structured APIs, not locked behind JavaScript rendering or authentication walls.
8. Create an llms.txt File
Provide AI crawlers with clear guidance about your brand, products, and positioning. This is the robots.txt equivalent for AI agents.
9. Monitor Your AI Visibility
Track how often AI agents recommend your brand, what they say about you, and where you're being missed.
When Agents Comparison Shop: Winning the Recommendation
Here's the scenario that matters: a consumer asks three different AI agents the same question. ChatGPT, Google AI Mode, and Perplexity each independently evaluate your brand against competitors.
In this multi-agent world:
- Agents don't respond to ads. ChatGPT explicitly serves organic-only results. You cannot buy placement.
- Agents evaluate structured data completeness. The brand with better Schema.org markup, more reviews, clearer pricing, and accessible product feeds wins.
- Agents cross-reference sources. If your product appears consistently across review sites, comparison platforms, and your own structured data, agents trust it more.
- Brands with incomplete data are invisible. Not ranked lower. Invisible. Agents can't recommend what they can't parse.
The 6.4x advantage for complete Schema.org markup means the gap between optimized and un-optimized brands is not gradual. It's binary. You're either in the conversation or you're not.
The winning strategy:
- Structured data completeness (table stakes)
- Review quality and volume (trust signals)
- Competitive pricing (comparison factor)
- API accessibility (discovery enabler)
- MCP connectivity (operational integration)
Frequently Asked Questions
How is agentic commerce different from regular AI chatbots?
Regular chatbots respond to questions. They answer "What's your return policy?" or "Is this in stock?" Agentic commerce involves AI agents that take independent action: discovering products across the internet, comparing options, making purchase decisions, and completing transactions. The agent pursues a goal, not just a conversation.
Do I need to rebuild my Shopify store for agentic commerce?
No. Agentic commerce optimization layers on top of your existing store. The primary requirements are structured data (Schema.org markup), quality product feeds, and optionally an MCP server connection. Your existing site, checkout flow, and operations stay intact.
Which AI shopping platform should I optimize for first?
Start with structured data and Schema.org markup because it benefits all platforms simultaneously. After that, prioritize based on where your customers are: ChatGPT has the largest user base (700M weekly users), Google AI Mode has the most product data (50B products in Shopping Graph), and Perplexity has the most purchase-intent traffic.
How long before agentic commerce becomes mainstream?
It already is for product research. 30-45% of US consumers use GenAI for product comparison today. Full autonomous purchasing is earlier stage, with 50% of consumers still cautious about letting AI complete purchases independently. The practical timeline: research and recommendations are mainstream now; autonomous purchasing will follow within 12-18 months as trust builds.
Get Agent-Ready Before Your Competitors Do
The brands capturing agentic commerce traffic today aren't waiting for the trend to mature. They're building structured data, connecting MCP servers, and monitoring their AI visibility while competitors still think "AI commerce" means adding a chatbot to their site.
$15 trillion in spending will flow through AI agents by 2028. The question isn't whether agentic commerce will reshape ecommerce. It's whether your store will be visible when it does.
See how Lexsis makes your brand discoverable to every AI shopping agent →
