AI in SAP Commerce and Marketing: Personalisation, Predictive Scores, Generative Content, Agents and Consent
As of 2026-11-05
As read on 5 November 2026, AI in SAP's commerce and marketing stack falls into three families: machine-learning scores and recommendations (Predict, AI segments, the CDP AI Workbench, Context-Driven Services), generative content (AI content composer, AI product finder, CX AI Toolkit tools) and agents (the Commerce MCP server, Joule in SAP Engagement Cloud). SAP renamed SAP Emarsys to SAP Engagement Cloud on 19 February 2026. This module reads the Data Usage FAQ into a feature register, explains the CDP and Commerce integrations with their consent rules, walks through MCP tool discovery and OAuth for agents, and says what the pages do not document. The lab builds a consent register and an MCP tool contract; the exercise is a design note for a two-brand retailer.
What you will learn
- Separate the AI features of SAP Commerce Cloud, SAP Engagement Cloud and SAP Customer Data Platform into scores and recommendations, generative content and agents
- Read the Data Usage FAQ of SAP Engagement Cloud into a register of features, models, data sent and owners
- Map consents between SAP Commerce Cloud, SAP Customer Data Platform and SAP Engagement Cloud with the naming rules and the synchronisation direction
- Describe how the Commerce MCP server discovers tools, authorises an agent and which tools you would leave out
- Quote announced assistants with the register wording instead of a date of your own
Concept overview
Answer first. AI in SAP's commerce and marketing stack is not one product. As read on 5 November 2026 it is three families of features living in three applications: machine-learning scores and recommendations (Predict, AI segments and send-time features in SAP Engagement Cloud, churn and lifetime-value models in SAP Customer Data Platform, the Recommendations feature of Context-Driven Services for SAP Commerce Cloud), generative content (the AI content composer, AI product finder and subject line generator in SAP Engagement Cloud, the AI tools of the SAP CX AI Toolkit for Commerce), and agents (an MCP server that lets external AI agents search, fill a cart and start checkout on a storefront, plus Joule in Engagement Cloud). Each family has a different data path, a different consent consequence and a different maturity, and the consultant's job is to keep them apart.
Prerequisites
- Basic knowledge of e-commerce and email marketing terms (catalogue feed, segment, opt-in)
- A first view of SAP Business AI and Joule; see M358 for Joule in Sales Cloud and Service Cloud
- For the lab: read access to the SAP Help pages cited, and optionally tenants of the three applications
Outcomes
- Explain which AI features use a model trained by SAP and which use a large language model, with the caveat that model names are volatile
- Produce a consent matrix with consent ids, direction of synchronisation and effect of withdrawal
- Write an `x-mcp-tool` block with name, prompt and `requiresAuth` and explain why the tool snapshot is fixed
- Name what the pages do not document: withdrawal handling for trained models, accuracy of scores, limits for agent orders
- Order the activation of AI features with an owner and a control group for each
Full module with the Consultant plan — €19.90/month. The full module adds: the decision framework · the end-to-end scenario walkthrough · the KPI scorecard · the anti-patterns · the code blocks · the knowledge check · the diagrams.
Open missions
120 live roles and contracts in SAP AI & Joule right now — see the SAP AI & Joule missions · freelance SAP contracts.