TL;DR
- 90% of B2B purchases will be intermediated by AI agents by 2028, representing $15 trillion in spending that brands must prepare for now
- AI shopping agents from ChatGPT, Google, and Perplexity are already making autonomous purchase decisions, skipping traditional search entirely
- Products with complete Schema.org markup are 6.4x more likely to be selected by AI agents than those without structured data
- Shopify's MCP Server and Universal Commerce Protocol (UCP) signal that agent-native commerce infrastructure is here, not coming
- The brands that win in agentic commerce are the ones optimizing for machine readability today, not just human browsing
- 64% of consumers plan to use AI chatbots for shopping in 2026, making agent optimization as critical as SEO was in 2010
What Is Agentic Commerce?
Agentic commerce is a new paradigm in ecommerce where AI agents act on behalf of consumers to research, compare, negotiate, and purchase products autonomously. Unlike traditional online shopping, where a human navigates a website, clicks through product pages, and completes a checkout form, agentic commerce delegates some or all of these steps to an AI agent that operates with real purchasing authority.
Think of it this way: traditional ecommerce is self-service. The consumer does the work. Agentic commerce flips that model. The consumer states an intent ("I need running shoes for trail running under $150 with good arch support"), and an AI agent handles the rest, from discovery through checkout.
This shift is not hypothetical. Shopping-related GenAI use grew 35% from February to November 2025 according to BCG research, and 30-45% of US consumers already use GenAI for product research according to Bain Consumer Lab. The trajectory is clear: consumers are delegating purchasing decisions to AI at an accelerating rate.
For brands, this means your next "customer" may not be a human browsing your site. It may be an AI agent evaluating your product data, comparing your specs against competitors, and making a purchase decision in milliseconds based on structured information rather than visual design or emotional marketing copy.
The implications are profound. When your buyer is an algorithm, the rules of engagement change completely.
The AI Shopping Agents Reshaping Ecommerce
Four major platforms have launched AI shopping agents in the past 18 months, each with a different approach but the same destination: autonomous purchasing on behalf of consumers.
ChatGPT Shopping
OpenAI launched shopping features within ChatGPT in April 2025, and the scale is staggering. With 700 million weekly active users, ChatGPT represents the largest deployment of AI shopping capabilities ever. When a user asks ChatGPT for product recommendations, it returns organic, unsponsored results based on its understanding of product quality, relevance, and fit.
This is critical for brands to understand. There are no paid placements in ChatGPT shopping results. You cannot buy your way to the top. Your product either earns a recommendation through genuine quality signals, structured data, and brand authority, or it does not appear at all.
Google AI Mode and "Buy For Me"
Google's AI Mode, powered by Gemini and the Shopping Graph (which indexes over 50 billion product listings), takes agentic commerce even further with its "Buy For Me" feature. This allows Google's AI agent to complete an entire purchase autonomously, navigating to a retailer's website, adding items to cart, entering payment information, and completing checkout without the consumer ever visiting the store.
This is the purest form of agentic commerce: a fully autonomous purchasing agent that eliminates the entire traditional ecommerce funnel. If your product is not surfaced in Google's Shopping Graph with complete, accurate structured data, it simply does not exist in this channel.
Perplexity Buy with Pro
Perplexity launched "Buy with Pro" in November 2024, enabling one-click checkout directly within its AI search interface. Users researching products can purchase without leaving the Perplexity environment. The agent handles product selection, comparison, and checkout in a single conversational flow.
What makes Perplexity significant is its research-first approach. Users come with complex purchase intent, and the AI agent synthesizes information across dozens of sources before recommending a product. This means your brand needs to be well-represented across the web, not just on your own domain.
OpenAI Operator
Launched in January 2025, OpenAI Operator is a vision-based AI agent that navigates websites exactly as a human would, but faster and without distraction. It can browse product pages, compare options, add items to cart, and complete purchases by visually interpreting web interfaces.
Operator represents a different approach to agentic commerce. Rather than relying on APIs or structured data feeds, it interacts with the visual layer of websites. This means site performance, clear navigation, and unambiguous product information become critical, because an AI agent reading your site has zero tolerance for confusing layouts or missing information.
From Search-to-Buy to Agent-to-Buy
The traditional ecommerce funnel has five steps: awareness, consideration, comparison, decision, and purchase. A consumer might spend days or weeks moving through these stages, visiting multiple sites, reading reviews, and comparing prices.
Agentic commerce collapses this funnel into a single interaction. A consumer tells their AI agent what they need, and the agent handles every subsequent step. The funnel becomes: intent, agent evaluation, purchase. Three steps instead of five, and only the first involves the human.
This compression has massive implications for brands. In the traditional funnel, you had multiple touchpoints to influence the buyer: ads for awareness, content for consideration, pricing for comparison, UX for conversion. In the agent-to-buy model, you get one shot. The AI agent evaluates your product data once, compares it against alternatives, and makes a decision.
Gartner projects that 90% of B2B buying will be AI-agent intermediated by 2028, representing $15 trillion in spending. On the consumer side, 64% of consumers plan to use AI chatbots for shopping in 2026 according to PartnerCentric research.
The brands that understand this shift and optimize for agent-readable commerce will capture disproportionate market share. The brands that continue optimizing only for human browsers will find themselves invisible to the fastest-growing purchase channel in history.
How AI Agents Discover and Choose Products
Understanding how AI shopping agents evaluate and select products is essential for any brand serious about AI visibility. The selection process differs fundamentally from how human shoppers make decisions.
Structured Data Is the New Storefront
AI agents do not browse. They parse. When an AI agent evaluates your product, it is not admiring your hero imagery or reading your brand story. It is extracting structured data: specifications, pricing, availability, reviews, shipping terms, and return policies.
Research from LLMRecommend.com in Q1 2026 found that products with complete Schema.org markup are 6.4x more likely to be selected by AI agents than products without structured data. This is the single most impactful finding for brands entering the agentic commerce era.
The Agent Decision Framework
AI shopping agents typically evaluate products across these dimensions:
- Relevance to stated intent - Does the product match what the consumer asked for?
- Specification completeness - Are all relevant attributes documented and machine-readable?
- Price-to-value ratio - How does pricing compare to alternatives with similar specifications?
- Availability and fulfillment - Can the product be delivered within the consumer's timeframe?
- Brand authority signals - What do reviews, citations, and third-party mentions indicate about quality?
- Return and warranty terms - What is the risk to the consumer if the product does not meet expectations?
Notice what is absent from this list: brand awareness, emotional connection, visual appeal, influencer endorsements, and loyalty programs. These factors, which dominate human purchasing, are largely irrelevant to an AI agent making a rational selection based on data.
Citation and Authority
AI agents weight brand authority heavily, but they measure it differently than humans do. Rather than brand recognition, agents look for consistent mentions across authoritative sources, expert reviews that reference specific product capabilities, and structured review data with high volume and recency.
This means your brand's presence across the web matters as much as your own site. Being cited in expert reviews, comparison articles, and industry publications feeds the authority signals that AI agents use for selection.
Shopify's Bet on Agent Commerce
Shopify's moves in 2025 and 2026 signal that the largest ecommerce platform in the world views agentic commerce as the next major channel. Two developments stand out.
Shopify Storefront MCP
Shopify's Model Context Protocol (MCP) Server allows AI agents to directly search products, create carts, and initiate checkout on any Shopify store. This is infrastructure-level support for agent commerce, meaning any AI agent that supports MCP can now transact on millions of Shopify stores without custom integration.
For merchants, this means your Shopify store is already accessible to AI agents if you have MCP enabled. The question is whether your product data is optimized for this channel.
Universal Commerce Protocol (UCP)
Shopify's Universal Commerce Protocol goes further, creating a standardized layer for AI agents to interact with commerce infrastructure. UCP treats AI agents as a first-class sales channel, equivalent to web, mobile, or social commerce.
The message from Shopify is clear: agent commerce is not an experiment. It is a production-ready channel that will grow to rival or exceed traditional web commerce in transaction volume. Brands that treat it as such will have a significant advantage.
How to Prepare Your Store for AI Agents
Preparing for agentic commerce requires a different mindset than traditional ecommerce optimization. Here is a practical checklist for brands ready to capture agent-driven revenue.
Product Data Foundation
- Complete Schema.org Product markup on every product page, including price, availability, SKU, brand, reviews, and specifications
- Machine-readable specifications with standardized units and formats (not buried in marketing copy or images)
- Structured review data with aggregate ratings, review count, and individual review markup
- Real-time inventory and pricing feeds that agents can verify (stale data kills trust)
- Clear, unambiguous product titles that include key specifications (agents parse titles for quick matching)
Technical Infrastructure
- Enable MCP on your Shopify store to allow direct agent interaction with your product catalog
- Maintain an updated product feed in Google Merchant Center (feeds the Shopping Graph used by Google AI Mode)
- Implement clean, fast-loading product pages (Operator and similar vision-based agents penalize slow or cluttered pages)
- Ensure your robots.txt allows AI crawlers including GPTBot, Google-Extended, PerplexityBot, and ClaudeBot
- Create an llms.txt file at your domain root summarizing your brand, products, and key differentiators for LLM consumption
Content and Authority
- Build citation-worthy content that AI agents can reference when explaining your products to consumers
- Earn mentions in expert reviews and comparison articles that AI agents use for authority scoring
- Maintain consistent NAP (name, address, phone) and brand information across all platforms
- Publish detailed FAQ content addressing common purchase objections in a structured, parseable format
- Keep product descriptions factual and specification-rich rather than purely emotional or aspirational
Monitoring and Optimization
- Track your AI visibility across ChatGPT, Gemini, Perplexity, and Claude to understand where agents recommend you (and where they do not)
- Monitor competitor visibility in AI agent recommendations
- Test agent interactions with your store by querying AI shopping agents for your product category
- Update structured data promptly when products change, because agents penalize stale or inaccurate information
When Agents Comparison Shop: Winning the Recommendation
The most critical moment in agentic commerce is when an AI agent comparison shops. The consumer has expressed intent, the agent has identified a shortlist of products, and now it must choose one. How do you win that moment?
Completeness Beats Creativity
In human ecommerce, creative product descriptions and emotional storytelling can differentiate your brand. In agent commerce, completeness wins. The brand with the most complete, accurate, and structured product data will be selected over the brand with better copywriting but sparse specs.
This does not mean abandoning creative marketing. It means ensuring that underneath your human-facing experience, there is a comprehensive machine-readable data layer that agents can parse and compare.
Price Transparency Matters More
AI agents are exceptional comparison shoppers. They can evaluate pricing across dozens of alternatives in milliseconds. Obscuring your pricing, hiding fees, or making total cost difficult to calculate will not trick an agent. It will simply cause the agent to select a competitor with transparent pricing.
Bold, clear pricing with all costs visible is table stakes for agent commerce.
Specification Standardization
When an AI agent compares running shoes, it needs to compare like with like: weight in grams, drop in millimeters, cushioning type by standard classification, and upper material. If your product uses non-standard specification formats or buries specs in paragraph text, the agent may not extract them properly, effectively making your product incomparable and therefore unselectable.
Standardize your product specifications using industry-standard terminology and units. Make them available in structured format, not just visual tables rendered as images.
Review Volume and Recency
AI agents heavily weight social proof, but they measure it quantitatively. A product with 2,000 reviews averaging 4.3 stars will typically be preferred over a product with 50 reviews averaging 4.8 stars. Volume indicates reliability of the signal.
Recency also matters. AI agents discount old reviews because product quality can change over time. A steady stream of recent positive reviews signals ongoing quality.
Availability and Fulfillment Speed
An AI agent making a purchase decision will factor in whether the product can actually be delivered within a reasonable timeframe. Products that are in stock, ship quickly, and have reliable delivery estimates will be preferred over products with ambiguous availability.
Ensure your inventory data, shipping estimates, and fulfillment capabilities are accurately represented in your structured data. Overpromising here creates negative signals when agents track fulfillment outcomes over time.
FAQ
How is agentic commerce different from conversational commerce?
Conversational commerce involves a human chatting with a bot to get product recommendations or support, but the human still makes the final purchase decision. Agentic commerce gives the AI agent actual purchasing authority. The agent does not just recommend products, it evaluates, selects, and buys them autonomously on the consumer's behalf. The distinction is agency: in conversational commerce the human decides, in agentic commerce the agent decides.
Will AI shopping agents replace traditional ecommerce websites?
Not entirely, but they will capture a growing share of transactions. Think of agent commerce as a new channel alongside web, mobile, and social. Some purchases (high-consideration, emotionally driven, or novel categories) will remain human-driven for years. But routine purchases, replenishment, and specification-driven buying (electronics, supplies, commodity goods) will shift rapidly to agents. Gartner's projection of 90% B2B agent intermediation by 2028 suggests the shift will be faster than most brands expect.
What should brands prioritize first for agent commerce readiness?
Start with structured data. Implementing complete Schema.org Product markup with all relevant properties (price, availability, reviews, specifications, brand, SKU) is the single highest-ROI action. Products with complete markup are 6.4x more likely to be selected by AI agents. After structured data, focus on enabling MCP access, building authority through expert citations, and monitoring your AI visibility across major platforms.
Does paid advertising work in agentic commerce?
Currently, the major AI shopping agents (ChatGPT, Perplexity) deliver organic, unsponsored recommendations. You cannot buy placement in ChatGPT shopping results. This may evolve as platforms monetize, but for now, earning recommendations through product quality, complete data, and brand authority is the only path. This makes agentic commerce fundamentally different from traditional search marketing, where paid placement can guarantee visibility.
The Agentic Commerce Opportunity Is Now
The window for early-mover advantage in agentic commerce is open, but closing. With $15 trillion in B2B spending moving to agent-intermediated channels by 2028, and consumer adoption accelerating at 35% growth rates, the brands that act now will establish the data foundations and authority signals that become exponentially harder to build later.
The playbook is clear: complete structured data, agent-accessible infrastructure, authority through citations, and continuous visibility monitoring. Brands that execute on these fundamentals will capture the next wave of commerce growth.
Lexsis helps ecommerce brands become visible, recommendable, and purchasable by AI agents. From AI visibility monitoring to agent-optimized storefronts, we give you the tools to win in the agentic commerce era. See how your brand performs across ChatGPT, Gemini, Perplexity, and Claude today.
