Quick navigation:
Shopping Online via AI Tools
Consumers are increasingly using AI tools and LLMs as the starting point for their online shopping journeys. In the past, they would start with a search engine (Google) and manually click through search results to find the right product. In 2026, this orientation and comparison process is largely handled by tools such as ChatGPT, Gemini, Claude, or Perplexity. As a result, AI is becoming increasingly important in e-commerce.
These LLMs provide suggestions and recommendations based on the questions users ask them. Instead of entering loose keywords, consumers now ask complete, contextual questions aligned with their needs and expectations, such as: “What is the best drill for a DIY beginner to drill into stone and concrete?”
This way of searching feels more natural, faster, and more personal. Consumers receive recommendations that perfectly match their query and budget. LLMs weigh different products against each other and present a curated selection of recommendations.
For now, consumers are often redirected to a webshop to complete their purchase, but even that step can be taken over by AI.
What Is Agentic Commerce?
Product purchases increasingly take place directly within AI interfaces such as ChatGPT, Gemini, and Copilot, without the consumer ever visiting the seller’s webshop. In the United States, these options are already widely available, and the rollout across Europe and other regions is progressing rapidly.
Completing payments within AI interfaces is an intermediate step toward the inevitable moment when consumers fully trust AI Agents to make purchases on their behalf. In that case, Agents will not only find and recommend products, but also order and pay for them.
When AI Agents can independently complete the entire process from orientation to transaction, this is called Agentic Commerce.
On-demand webinar
Watch on-demand: “Agentic AI in Ecommerce –
How LLMs Understand and Sell Your Products”
In this 30-minute webinar, you will learn more about Agentic Commerce and discover how to optimize and structure your product data so that LLMs can correctly interpret, compare, and recommend your products.
GEO: Optimize Your Product Data for AI
For retailers and e-commerce businesses, these developments mean that LLMs are no longer just an orientation channel, they are becoming an important sales channel. When ChatGPT, Perplexity, Claude, or Gemini presents your product as the best option, the likelihood of conversion is high.
Visibility and relevance are crucial. The right product must be recommended at the right moment and in the right way. It’s therefore becoming increasingly important to structure your product data in such a way that AI tools understand who a product is for, what its characteristics are, in which situations it is used, and what possible alternatives exist.
Optimizing your website for AI is known as Generative Engine Optimization (GEO), an extension of traditional Search Engine Optimization (SEO).
How Can AI Tools Understand Your Products?
An AI does not select a product as a recommendation simply because it is available, but because it understands what problem the product solves and when it is relevant to a user. LLMs connect the context and preferences provided in a user’s query to product information.
The more complete your product data is, and the more clearly it is structured, the greater the chance that your products will be recommended. Product data is therefore no longer just input for product pages or SEO, but also for AI tools acting as advisors and comparison engines. Context, consistency, and semantics play a crucial role.
Optimizing product content is no longer just about being found through keywords (SEO). It is about being understood by answering the questions your potential customers have (GEO). Rich context descriptions are essential. Read more about GEO
Key Product Data Elements for Agentic Commerce
Which aspects of product data are important for Agentic Commerce?
- Categories & taxonomy
A clear category structure provides LLMs with context and makes relationships between products understandable. - Attributes & specifications
Factual specifications form the foundation for comparison and targeted recommendations. - Titles & descriptions
Clear product texts add context and meaning. Who is the product for? When is it used? What differentiates it from alternatives? See also: GEO product content - Images & visual metadata
For LLMs that can process not only text but also images and other media, high-quality visuals and metadata help them better recognize and interpret products.
GEO Step-by-step guide
GEO: How to Optimize Product Data for AI-Driven Shopping.
How Do I Sell Products via ChatGPT, Copilot, and Gemini?
Rich product data will remain essential for visibility and discoverability in the future, but Agentic Commerce goes further than that.
Products can be purchased directly within AI interfaces, or paid for autonomously by AI Agents. If you want consumers to order products directly within these AI environments, your webshop must integrate with the UCP and/or ACP protocol.
ACP (Agent Communication Protocol)
Developed by OpenAI and Microsoft, ACP is essential for AI commerce and checkout functionality within ChatGPT and Copilot. ACP governs how AI agents communicate, collaborate, and make joint purchasing decisions.
UCP (Universal Commerce Protocol)
Developed by Google, UCP is necessary for AI commerce and checkout options within Google Gemini and Google Search. UCP ensures that AI agents can search, compare, and purchase products across platforms while interpreting product data uniformly.
Implementing ACP/UCP for Agentic Commerce
To enable AI Agents to find and purchase your products, you need a product feed that complies with UCP and/or ACP standards.
This feed contains all essential product data, such as prices, stock levels, images, descriptions, accessories, and return policies. AI Agents use this feed as a source to search products, check availability, and process orders in real time.
Good to know: If you use Shopify as your e-commerce platform, ACP/UCP support is already built in. This means you are automatically prepared for AI commerce.
If you are not using Shopify, you must meet several requirements:
Google (Gemini) – UCP requirements:
- Active Google Merchant Center account
- Complete and up-to-date product feeds including brand assets, accessories, and media
- Clearly defined return policy with costs, timelines, and link to full terms
- Customer support and company contact information
OpenAI (ChatGPT) / Microsoft (Copilot) – ACP requirements:
- Active ChatGPT Merchant account
- Secure, up-to-date product feed (CSV or JSON) with identifiers, descriptions, prices, stock levels, media, and fulfillment options
- Rich media, reviews, FAQs, accessories, replacement options, and cross-sell products to improve ranking, relevance, and recommendations
Optimize Your Product Data for Agentic Commerce
Agentic Commerce fundamentally changes how consumers find and purchase your products. AI tools are taking over more and more steps in the buying journey, from orientation and comparison to actual transaction.
For retailers and e-commerce businesses, this means product data must not only be optimized for search engines and product pages, but also be understandable and actionable for AI Agents.
Organizations that invest in structured, complete, and context-rich product information ( and are ready for ACP and UCP integrations), ensure that AI Agents can correctly interpret, recommend, and sell their products.
Live demo
