[et_pb_section fb_built=”1″ admin_label=”Hero” _builder_version=”4.27.5″ _module_preset=”default” custom_padding=”0vw||80px||false|false” locked=”off” collapsed=”on” global_colors_info=”{}”][et_pb_row _builder_version=”4.18.0″ _module_preset=”5138c454-be54-4233-bd3b-f8e6a8747976″ global_colors_info=”{}”][et_pb_column type=”4_4″ _builder_version=”4.18.0″ _module_preset=”73121f80-a3ef-4484-8763-c3f18e3c56d2″ global_colors_info=”{}”][et_pb_heading title=”SAP Commerce” _builder_version=”4.27.5″ _module_preset=”f0c675ea-2574-4d0e-b725-30f8550a8550″ title_level=”h4″ title_font=”IBM Plex Sans|IBM Plex Sans Condensed_weight||on|||||” title_text_color=”#1d3557″ title_font_size=”14px” title_letter_spacing=”1px” title_line_height=”1.4em” custom_margin=”||10px||false|false” locked=”off” global_colors_info=”{%22#1d3557%22:%91%22title_text_color%22%93}” theme_builder_area=”post_content”][\/et_pb_heading][et_pb_heading title=”Agentic AI and MCP Server for SAP Commerce: What It Means for Your Storefront” _builder_version=”4.27.5″ _module_preset=”ddeb8c09-9078-4424-bc15-2efb6572e28e” title_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}”][\/et_pb_heading][et_pb_post_title title=”off” meta=”off” featured_image=”on” _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}”][\/et_pb_post_title][et_pb_text quote_border_color=”#457b9d” _builder_version=”4.27.5″ _module_preset=”2c55a9c4-feed-423b-9edb-ae0b5b365cac” text_font=”IBM Plex Sans|IBM Plex Sans_weight|||||||” link_font=”|IBM Plex Sans_weight|||||||” link_text_color=”#1d3557″ quote_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_2_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_3_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_4_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_5_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_6_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}” theme_builder_area=”post_content”]<\/p>\n
# Agentic AI and MCP Server for SAP Commerce: What It Means for Your Storefront\n<\/p>\n
At NRF 2026 in January, SAP made an announcement that did not get the attention it deserves. Buried among the usual wave of retail technology news, SAP revealed a storefront MCP (Model Context Protocol) server for SAP Commerce Cloud, with general availability planned for Q2 2026. If you are running SAP Commerce and your leadership team has not discussed this yet, you are already behind on a conversation that will reshape how customers find and purchase your products.\n<\/p>\n
This is not another chatbot bolted onto a product page. This is about AI agents – from Claude, ChatGPT, Perplexity, and others – acting on behalf of consumers to browse, compare, and purchase products across the internet. Your storefront is either accessible to these agents, or it is invisible to them. There is no middle ground.\n<\/p>\n
Let’s break down what happened, what it means in practice, and what you should actually do about it.\n<\/p>\n
[\/et_pb_text][et_pb_divider _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_divider][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_column type=”4_4″ _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_heading title=”What Is the Model Context Protocol (MCP)?” _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_heading][et_pb_text quote_border_color=”#457b9d” _builder_version=”4.27.5″ _module_preset=”2c55a9c4-feed-423b-9edb-ae0b5b365cac” text_font=”IBM Plex Sans|IBM Plex Sans_weight|||||||” link_font=”|IBM Plex Sans_weight|||||||” link_text_color=”#1d3557″ quote_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_2_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_3_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_4_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_5_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_6_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}” theme_builder_area=”post_content”]<\/p>\n
Before diving into SAP’s implementation, you need to understand MCP itself. The Model Context Protocol is an open standard developed by Anthropic (the company behind Claude) that defines how AI agents interact with external systems. Think of it as a structured API specifically designed for AI consumption, not human-facing interfaces.\n<\/p>\n
Traditional web APIs were designed for developer-built applications: a frontend calls an endpoint, parses the response, and renders it for a human. MCP flips this model. An AI agent calls the MCP server, receives structured data it can reason about, and takes actions on behalf of a user. The agent understands context, handles multi-step workflows, and can make decisions that previously required a human clicking through pages.\n<\/p>\n
The protocol is open, meaning any AI system can implement it. This is not a proprietary SAP-only standard. When SAP builds an MCP server for Commerce Cloud, it means that Claude, ChatGPT, Perplexity, and any other MCP-compatible agent can interact with your storefront programmatically.\n<\/p>\n
You might be thinking: “We already have REST APIs and OCC (Omni Commerce Connect) endpoints. How is this different?”\n<\/p>\n
The difference is intent and audience. Your OCC APIs serve your own frontend applications. You control the client, the UX flow, and the data presentation. MCP serves external AI agents that you do not control. These agents have their own reasoning capabilities, their own user relationships, and their own decision frameworks. Your storefront becomes one of many data sources an agent queries when a consumer says, “Find me the best industrial pump under 5,000 EUR with next-day delivery.”\n<\/p>\n
If your competitor’s storefront answers that query and yours does not, the agent will never mention your products.\n<\/p>\n
[\/et_pb_text][et_pb_divider _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_divider][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_column type=”4_4″ _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_heading title=”What SAP’s MCP Server Actually Does” _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_heading][et_pb_text quote_border_color=”#457b9d” _builder_version=”4.27.5″ _module_preset=”2c55a9c4-feed-423b-9edb-ae0b5b365cac” text_font=”IBM Plex Sans|IBM Plex Sans_weight|||||||” link_font=”|IBM Plex Sans_weight|||||||” link_text_color=”#1d3557″ quote_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_2_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_3_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_4_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_5_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_6_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}” theme_builder_area=”post_content”]<\/p>\n
Based on what SAP has disclosed and demonstrated, the Commerce Cloud MCP server exposes several core capabilities to AI agents:\n<\/p>\n
Product discovery and search.<\/strong> Agents can browse your product catalog, filter by attributes, search by natural language queries, and retrieve detailed product information including specifications, pricing, and availability. This goes beyond simple keyword matching. The agent understands product categories, can compare alternatives, and can reason about which products best match a user’s stated needs.\n<\/p>\n Cart and checkout operations.<\/strong> Agents can add products to a cart, apply promotions, calculate shipping, and initiate the purchase flow. The exact scope of checkout completion is still being defined (payment authorization through an AI agent raises significant UX and security questions), but the product selection and cart management capabilities are confirmed.\n<\/p>\n Inventory and availability checks.<\/strong> Real-time stock levels and delivery estimates, which are critical for agents making recommendations. An agent that recommends an out-of-stock product loses user trust quickly.\n<\/p>\n Account-aware interactions.<\/strong> For B2B scenarios especially, the MCP server can respect customer-specific pricing, contract terms, and catalog restrictions. This is where SAP’s enterprise capabilities give it an edge over simpler MCP implementations.\n<\/p>\n Let’s be honest about the current limitations. The MCP server in its initial release will not handle:\n<\/p>\n These limitations matter. If your business depends heavily on CPQ or complex approval chains, the MCP server alone will not replace your existing customer-facing channels. It is an additional channel, not a replacement.\n<\/p>\n [\/et_pb_text][et_pb_divider _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_divider][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_column type=”4_4″ _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_heading title=”SAP’s Broader Agentic Commerce Strategy” _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_heading][et_pb_text quote_border_color=”#457b9d” _builder_version=”4.27.5″ _module_preset=”2c55a9c4-feed-423b-9edb-ae0b5b365cac” text_font=”IBM Plex Sans|IBM Plex Sans_weight|||||||” link_font=”|IBM Plex Sans_weight|||||||” link_text_color=”#1d3557″ quote_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_2_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_3_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_4_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_5_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_6_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}” theme_builder_area=”post_content”]<\/p>\n The MCP server is one piece of a larger strategy SAP is building around what they call “agentic commerce.” At NRF and in subsequent briefings, SAP outlined several AI agents designed to work alongside the MCP server:\n<\/p>\n SAP claims this agent can analyze your product data and suggest improvements to titles, descriptions, attributes, and categorization, with a stated target of 70% faster content improvement cycles. In practice, this means the agent reviews your product information against best practices for AI discoverability and flags gaps. Missing attributes that an external AI agent would need, inconsistent naming conventions, sparse descriptions that limit an agent’s ability to match products to queries.\n<\/p>\n The 70% number is a marketing claim, but the underlying capability is real and valuable. Product data quality has always mattered for search and conversion. It matters even more when the “searcher” is an AI agent that reasons about structured data rather than scanning a webpage visually.\n<\/p>\n SAP has announced promotion-related AI capabilities that manage promotion targeting based on customer behavior, segments, and purchase patterns. In concept, this automates work that merchandising teams currently do manually in the Backoffice: creating promotion rules, setting audience segments, scheduling campaigns. The system observes what is working, adjusts targeting, and suggests new promotions.\n<\/p>\n The practical value depends heavily on your current promotion complexity. If you run a handful of straightforward discounts, the AI adds marginal value. If you manage hundreds of customer-segment-specific promotions across multiple markets (as we have seen in multi-country telecom and automotive implementations), the automation potential is significant.\n<\/p>\n A customer support agent that can handle post-purchase inquiries: order status, delivery tracking, return initiation, and basic troubleshooting. This is the most mature category of AI agents in e-commerce, and SAP’s implementation builds on their existing SAP Service Cloud capabilities.\n<\/p>\n Announced as part of SAP Order Management Services, this agent focuses on order fulfillment reliability and prediction. It helps identify potential fulfillment issues before they affect the customer, improving delivery accuracy and proactively managing exceptions in the order lifecycle.\n<\/p>\n [\/et_pb_text][et_pb_divider _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_divider][\/et_pb_column][\/et_pb_row][et_pb_row _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_column type=”4_4″ _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][et_pb_heading title=”What This Means for Your Technology Roadmap” _builder_version=”4.27.5″ _module_preset=”default” global_colors_info=”{}” theme_builder_area=”post_content”][\/et_pb_heading][et_pb_text quote_border_color=”#457b9d” _builder_version=”4.27.5″ _module_preset=”2c55a9c4-feed-423b-9edb-ae0b5b365cac” text_font=”IBM Plex Sans|IBM Plex Sans_weight|||||||” link_font=”|IBM Plex Sans_weight|||||||” link_text_color=”#1d3557″ quote_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_2_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_3_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_4_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_5_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” header_6_font=”IBM Plex Sans Condensed|IBM Plex Sans Condensed_weight|||||||” locked=”off” global_colors_info=”{}” theme_builder_area=”post_content”]<\/p>\n This is where theory meets practice. Here is what you should actually be evaluating and planning for.\n<\/p>\n SAP’s MCP server is built to work with the Composable Storefront (Spartacus) and the underlying OCC APIs. If you are still running an Accelerator-based storefront, your path to MCP integration adds a prerequisite step: you need to be on (or migrating to) the Composable Storefront first.\n<\/p>\n This does not mean you need to drop everything and rewrite your storefront. But if a Spartacus migration was already on your roadmap for the next 12 to 18 months, the MCP server adds weight to that decision. The longer you stay on Accelerator, the further you fall behind on capabilities that require a modern, API-first architecture.\n<\/p>\nWhat It Does Not Do (Yet)<\/h3>\n
\n
Catalog Optimization Agent<\/h3>\n
AI-Powered Promotion Optimization<\/h3>\n
Digital Service Agent<\/h3>\n
Order Reliability Agent<\/h3>\n
1. Composable Storefront Readiness<\/h3>\n