Since July 2026, buyers can dispatch digital assistants that independently research products, compare prices and trigger orders. This shift toward "agentic commerce" is fundamentally changing the rules in B2B sales. Shopware responded with version 6.7.12.1 and introduced the "AI Discovery Files per sales channel" feature. But what does this mean in practice for manufacturers and wholesalers? And how do you ensure that AI agents find and understand your shop in the first place?
This article explains why AI Discovery Files are relevant for B2B shops, how the new Shopware feature works and which five concrete steps you should implement now so that your digital sales channel stays competitive in the era of autonomous purchasing agents.
What is agentic commerce and why does it matter for B2B?
Agentic commerce describes the use of AI-powered agents that autonomously carry out procurement processes on behalf of companies or buyers. Unlike classic search engines or chatbots, these agents analyze structured and unstructured data, compare offers across multiple platforms and make purchasing decisions based on predefined criteria such as price, delivery time, quality or sustainability. How AI agents are used sensibly in B2B is increasingly evident in procurement as well.
For B2B companies this means: customers no longer expect just a functioning online shop, but a machine-readable, structured data foundation. Anyone who does not prepare their product information, prices and availability in a way that AI agents can understand and process will simply not be considered in automated procurement processes. In parallel, the Shopware Copilot is becoming increasingly agentic and takes over tasks within the shop itself.
The challenge lies less in the technology itself and more in data quality. Many mid-sized manufacturers and wholesalers have stored their product data in ERP systems, Excel spreadsheets and PDF catalogs over the years. This data is readable for humans but difficult for machines to interpret. This is exactly where AI Discovery Files come in.
What are AI Discovery Files and how do they work?
AI Discovery Files are structured files that provide information about your shop, your assortment and your terms of business in a machine-readable format. They work similarly to a sitemap for search engines, but are specifically optimized for AI agents.
With version 6.7.12.1, Shopware introduced native support for AI Discovery Files. The feature automatically creates a JSON file per sales channel that is available at a defined URL. Among other things, this file contains:
Basic shop information (name, industry, target group)
Product categories and their structure
Available payment and shipping options
Contact information and support channels
Links to further machine-readable data sources (e.g. product feeds)
AI agents can read this file and decide within seconds whether your shop is relevant for a specific request. If this file is missing or incomplete, the probability that your offer is considered at all drops considerably.
The introduction of this feature is not a marketing gimmick, but a response to a real market trend. Large purchasing platforms and procurement systems already rely on AI-supported processes. Mid-sized B2B providers who act now secure an advantage over competitors who wait.
Step 1: Activate and configure the AI Discovery File in Shopware
The first step is technically manageable but requires a conscious decision. In Shopware 6.7.12.1 and higher, you will find the "AI Discovery Files" feature in the settings of your sales channel. Activate the feature and check which information is filled in automatically.
Shopware pulls a lot of data from existing configurations: shop name, categories, payment methods, shipping methods. Nevertheless, you should manually review and complete the generated file. Typical weak points are:
Missing or inaccurate descriptions of product categories
Incomplete information about target groups or industries
Missing links to structured data sources such as XML feeds or API endpoints
Take the time to design the file so that an AI agent understands at a glance what you offer, for whom and under what conditions. A well-configured AI Discovery File is your digital business card for autonomous systems.
Do not forget to store the URL of the AI Discovery File in your robots.txt and, if necessary, register it in industry-specific directories. The more visible the file, the higher the chance that AI agents will find it.
Step 2: Structure and enrich product data
The AI Discovery File is only as good as the data it points to. Many B2B shops fail not because of the technology, but because of incomplete, inconsistent or simply missing product information. AI agents need structured data to be able to compare and evaluate products. Which data gaps in agentic commerce occur particularly often is worth a closer look.
In concrete terms, this means:
Complete product attributes: Every product should contain at least an item number, name, description, price, availability, unit, weight and dimensions. For technical products, standards, certificates and material properties are also relevant.
Categorization according to standards: Use established classification systems such as eCl@ss, UNSPSC or GS1. These standards make it considerably easier for AI agents to assign your products to search queries.
Semantic enrichment: Supplement your product data with synonyms, application examples and areas of use. A "screwdriver" should also be findable as a "screw turner." An "industrial hose" should be linked to use cases such as "food industry" or "chemical-resistant."
Images and documents: AI agents can increasingly process visual information as well. High-quality product images, technical drawings and data sheets should be stored in machine-readable formats (e.g. PDF/A with OCR).
If you do not yet use a Product Information Management (PIM) system, now is the right time to consider it. A PIM helps you manage, version and distribute product data centrally across channels. When a PIM rollout in mid-sized companies really pays off depends on assortment breadth and the number of channels. For mid-sized companies there are now also lean, affordable solutions that integrate seamlessly into Shopware.
Step 3: Make prices, availability and conditions transparent
AI agents make purchasing decisions not emotionally, but on the basis of hard facts. Price, delivery time and conditions are the most important decision criteria. Anyone who does not provide clear, up-to-date information here falls off the radar.
Make sure that your prices are synchronized in real time from the ERP system. Outdated prices lead to complaints, returns and loss of trust. Why real-time prices in the B2B shop decide between order and abandonment is already evident today. The same applies to availability: an AI agent that hits "not deliverable" three times will ignore your shop next time.
For B2B customers, customer-specific prices, graduated prices and framework agreements are also relevant. Shopware offers flexible options here that you should definitely use. AI agents can read this information and include it in their decision-making, provided it is stored in a structured way.
Delivery and payment terms should also be communicated clearly. Do you offer different shipping options? Are there minimum order quantities? Which payment terms apply? All this information belongs in the AI Discovery File and should be retrievable at the product level.
Transparency creates trust. And trust is a decisive competitive advantage in agentic commerce as well.
Step 4: Provide API interfaces for real-time communication
AI Discovery Files are a first point of contact. For deeper interactions, however, AI agents need direct interfaces to your system. Shopware offers a powerful API for this, enabling agents to retrieve product information in real time, check prices, create carts and trigger orders.
The Shopware Store API is available and documented by default. However, you should check whether all relevant endpoints are activated and secured. Typical use cases for AI agents are:
Product search and filtering by attributes
Retrieval of prices and availability
Creation and management of carts
Authentication and access to customer-specific conditions
Order processing and status queries
For many mid-sized companies, the idea of letting external systems access their shop API directly is initially unfamiliar. But this is exactly where the future of B2B sales lies. Anyone who opens their systems and makes them machine-readable gains reach and efficiency.
Pay attention to security: API access should be authenticated via OAuth2 or comparable procedures. Rate limiting and monitoring help prevent misuse and ensure performance.
Step 5: Maintain, test and continuously optimize data
Setting up AI Discovery Files and API interfaces is not a one-time project, but a continuous process. Product data changes, new items are added, prices are adjusted. Only if your data is current and correct will you remain relevant for AI agents.
Establish clear responsibilities for data maintenance. In many companies it is unclear who is responsible for product data: marketing, sales, IT or purchasing? A functioning data governance model creates clarity and ensures that data is not just captured once, but maintained permanently.
Regularly test how your shop looks from the perspective of an AI agent. There are now tools and simulators that make exactly this possible. Use them to identify and fix weak points.
In addition, analyze which requests from AI agents reach you and how they convert. Shopware offers extensive reporting functions for this. Use this data to continuously sharpen your strategy.
Agentic commerce is not hype, but a structural change in B2B sales. Companies that lay the foundations now will benefit in the long term.
How AI Discovery Files fit into your platform strategy
The introduction of AI Discovery Files is more than a technical adjustment. It is part of a broader question: how do I position myself as a manufacturer or wholesaler in an increasingly platform-based B2B world?
Many companies face the decision of whether to rely on their own platforms, go to existing B2B marketplaces or pursue hybrid models. AI Discovery Files are relevant in all three scenarios:
Own platform: Anyone running their own Shopware shop creates, with AI Discovery Files, the prerequisite for it to be found and used by external systems as well. Data sovereignty is preserved, reach is increased.
B2B marketplaces: Structured data also plays a central role on marketplaces such as Amazon Business, Mercateo or industry-specific platforms. Maintaining this data in your own shop considerably eases the connection.
Hybrid models: Many companies rely on multi-channel strategies. AI Discovery Files make it possible to use the same data structures across channels, thereby increasing efficiency and consistency.
The question is not whether you should rely on AI-supported processes, but how quickly you create the necessary foundations. Those who invest today secure competitive advantages tomorrow.
