Agentic Commerce Protocol (ACP)
The Agentic Commerce Protocol (ACP) is one of the technologies enabling the next generation of AI-driven commerce. It provides a framework for how AI agents can interact with commerce and payment systems to support product discovery, purchasing, and post-purchase activities on behalf of buyers.
For B2B organizations, ACP opens new opportunities to deliver AI-powered buying experiences while building on existing commerce, procurement, and ERP systems.
This guide explores the role of ACP in modern B2B and enterprise commerce and how it supports broader agentic commerce strategies.
What is ACP?
The Agentic Commerce Protocol (ACP) is an open standard and interaction framework, co-developed by OpenAI and Stripe, that enables AI agents to securely connect with merchant and payment systems and complete purchases on behalf of users.
With ACP, businesses implement a standardized interface to expose product catalogs, pricing, availability, and checkout capabilities. AI agents such as ChatGPT can then discover products, compare options, configure purchases, and initiate transactions directly within a conversational interface.
Importantly, ACP does not replace existing commerce infrastructure. Merchants remain the merchant of record and continue to use their established e-commerce, ERP, order management, payment, and fulfillment systems. ACP simply provides a common protocol that allows AI agents to interact with these systems in a secure and standardized way.
Is ACP only for B2C commerce?
No. Although many early ACP examples focus on consumer shopping experiences, the protocol is designed as a flexible and extensible standard that can support both B2C and B2B commerce.
In enterprise environments, ACP can facilitate complex purchasing processes, including procurement workflows, account-based purchasing, negotiated pricing, approval chains, contract-specific catalogs, and multi-user buying scenarios. This makes ACP relevant not only for online retail but also for manufacturers, distributors, wholesalers, and other B2B organizations looking to enable AI-assisted purchasing.
How is ACP different from a traditional online shop?
ACP does not replace a traditional online store. Instead, it standardizes how AI agents interact with commerce systems throughout the purchasing process.
In a conventional e-commerce experience, buyers navigate a website, search for products, compare options, and complete checkout themselves. With ACP, these interactions can be performed by an AI agent on the user's behalf through a conversational interface.
While the storefront may become less visible to the end user, the underlying commerce platform remains responsible for managing product catalogs, pricing, inventory, business rules, payments, orders, and fulfillment. ACP acts as the communication layer that enables AI agents and commerce systems to work together seamlessly.
Benefits of ACP for B2B companies
1. AI-powered buying experiences without rebuilding your stack
ACP enables enterprises to connect AI agents to existing commerce, ERP, and payment systems. Buyers can research, configure, and order products through natural language interactions, while businesses retain their established processes, integrations, and data.
2. Single integration, multiple AI channels
ACP provides a standardized layer between merchants, AI agents, and payment providers. Instead of building custom integrations for every platform, businesses can integrate once and support transactions across multiple AI-driven channels.
3. Merchant of record and business control
Even when a purchase is initiated through an AI agent, the business remains the merchant of record. It controls product availability, pricing, terms, and fulfillment processes, helping ensure compliance, governance, and contractual control — especially important in B2B environments.
4. Better support for complex B2B buying journeys
ACP supports structured product data, configuration workflows, and validation rules. AI agents can guide buyers through compatibility checks, contract-specific pricing, approval processes, and other requirements before triggering a compliant order in backend systems.
5. A foundation for agentic commerce strategies
ACP supports the broader vision of Agentic Commerce, where AI agents assist with product discovery, evaluation, and purchasing on behalf of buyers. Combined with AI-powered self-service portals and digital assistants, ACP can become a key building block of a modern agentic commerce strategy.
6. Future-proofing for AI-driven buying behavior
As AI agents increasingly influence how products are discovered and purchased, ACP provides a structured way to keep catalogs, pricing, and offers accessible through emerging AI channels — without abandoning proven omnichannel, composable, or headless commerce architectures.
Challenges & solutions
Challenge 1: Fragmented system landscapes
Many B2B organizations run multiple ERPs, legacy commerce platforms and custom portals. Exposing coherent ACP endpoints on top of this complexity can be difficult.
- Solution: Composable commerce and API-first design
An API-first, composable commerce platform decouples frontends and channels from backend systems. Commerce APIs, product services and order management can expose consistent interfaces that are easier to map to ACP actions.
Challenge 2: Data quality and product semantics
AI agents need structured, consistent product data, attributes and pricing rules to recommend and configure items safely.
- Solution: Centralized product and content services
With centralized product information, search and content services, enterprises can enrich catalog data, define semantic attributes and expose ACP-ready product representations. This also supports AI-driven self-service portals and agentic commerce scenarios.
Challenge 3: Governance, compliance and risk
In B2B, incorrect orders, wrong configurations or non-compliant deliveries due to AI agents can be costly and risky.
- Solution: Policy-aware order workflows
Commerce-based solutions can enforce approval flows, contract rules, credit limits and regional restrictions even when orders originate from ACP agents. Policies run in the commerce and order management layer, ensuring that only valid transactions get accepted.
Challenge 4: Integrating ACP into existing customer journeys
Sales teams, key account managers and procurement portals already serve customers. Adding ACP should not create silos or channel conflicts.
- Solution: Unified customer and order view
By using the commerce platform as a hub, all orders — whether from classic webshops, EDI, marketplaces or ACP — end up in one unified order management and customer history. This gives sales and service teams full visibility and avoids fragmented experience.
Challenge 5: Keeping AI agents aligned with brand and pricing strategy
AI agents might optimize for speed, not necessarily for margin, brand positioning or strategic products.
- Solution: Curated offers and configuration rules
Through APIs and business logic, enterprises can define which products and assortments are available via ACP, how they are presented and which configuration shortcuts are allowed. Rule engines, segmentation and personalization capabilities can inform these choices.
How ACP Works
ACP defines a standardized framework for interactions between three key participants: the buyer, the AI agent acting on the buyer’s behalf, and the merchant and payment systems that execute the transaction.
In a typical workflow, a buyer expresses a need in natural language. The AI agent then discovers relevant products through ACP-compatible interfaces, evaluates available options, and configures the purchase according to the buyer’s requirements. Next, the agent creates a cart, coordinates the checkout process, and initiates the transaction within the merchant’s existing commerce and payment systems.
After the purchase, the same protocol can be used to access order information, track order status, manage returns, or support reordering, creating a seamless end-to-end buying experience.
ACP & Intershop
With the Spring 2026 Release, the Intershop Commerce Platform natively supports ACP product feeds, enabling B2B merchants to make their catalogs discoverable in LLM-based environments such as ChatGPT — with no separate integration required.
ACP use cases
Technical spare parts procurement
A maintenance engineer describes a problem in natural language. An AI agent uses ACP to identify compatible spare parts based on machine models and serial numbers, checks availability and contract pricing, and places an order through existing B2B commerce and ERP systems.
Complex product configurations and BOMs
Manufacturing buyers often require configured products or complete bills of materials (BOMs). AI agents can ask clarifying questions, guide users through configuration options, and generate structured orders via ACP, while backend systems validate compatibility, pricing, and business rules.
Recurring B2B orders and replenishment
Purchasing teams can instruct an AI agent to reorder consumables or frequently purchased items with specific adjustments. The agent reviews order history, creates the order through ACP, and submits it to the relevant commerce and ERP systems.
AI-driven self-service portals
Existing customer portals remain in place, but AI agents become the primary user interface. Instead of navigating categories or search filters, buyers simply describe their needs, while the agent handles product discovery, configuration, and checkout through ACP.
Multi-brand and marketplace scenarios
In multi-brand, distributor, or marketplace environments, ACP can route orders from a single AI interaction to multiple merchants. Each merchant retains control over its own catalog, pricing, inventory, and fulfillment processes while participating in a seamless buying experience.
Frequently asked questions about ACP
Who maintains the Agentic Commerce Protocol (ACP)?
The Agentic Commerce Protocol (ACP) is currently maintained by OpenAI and Stripe in collaboration with commerce, payments, and technology partners. As an open standard, ACP is designed to support interoperability between AI agents, merchants, and payment providers across different platforms and ecosystems.
How does ACP relate to Model Context Protocol (MCP)?
The Model Context Protocol (MCP) enables AI agents to access tools, systems, and structured data. ACP builds on these capabilities by defining how AI agents interact with commerce and payment systems to discover products, manage carts, and complete transactions.
In simple terms, MCP helps agents understand and access information, while ACP enables them to perform commerce-related actions
Do I need a new shop system to support ACP?
No. ACP is designed to work alongside existing commerce, payment, and backend systems rather than replace them. Businesses can expose products, pricing, and checkout capabilities through ACP while continuing to use their current infrastructure. However, organizations with API-first, composable, or headless commerce architectures may find ACP integration easier to implement and scale.
Prepare your business for agentic commerce
AI agents are changing how buyers discover, evaluate, and purchase products. Whether you're exploring ACP adoption or building a broader agentic commerce strategy, Intershop helps you connect AI-driven buying experiences with your existing commerce infrastructure. Contact us to learn how to make your products and services ready for the next generation of digital commerce.
