SAP aligns commercial data structures for operational AI personalization

SAP has announced an initiative to align fragmented commercial data structures—so that AI personalization works not in theory, but at the level of operational execution.

The Gap Between Strategy and Reality

Large companies declare strategies to "anticipate customer needs" and build relevant experiences at every digital touchpoint. The problem is that the infrastructure meant to deliver this is structured differently. Customer data, product catalogs, transaction history, and behavioral analytics are stored in separate systems with incompatible schemas. Recommendation engines produce mundane listings not because the algorithms are poor, but because they're fed disconnected data.

SAP calls this the execution layer problem: the gap between strategic goals and what physically happens at the moment of customer interaction. Management sets KPIs for personalization, data teams build pipelines, but in the end the customer sees "we recommend" with products they've already bought.

What Exactly SAP Is Changing

The initiative is aimed at standardizing how commercial data is structured and interrelated. The goal is for the AI layer to operate on top of a unified semantic base, rather than attempting to interpret each system separately.

Key areas of change:

• Data schema unification — aligning formats between SAP Commerce Cloud, SAP Customer Data Platform, and related solutions into a single model
• Real-time linking — transactional data and user behavior are unified into a context accessible to the AI engine directly at query time
• Execution-layer API — interfaces through which personalization is embedded in commercial processes, rather than remaining in analytical dashboards
• Reduced manual mappings — fewer ETL pipelines that break whenever any system is updated
• LLM compatibility — data structures are adapted to work with language models without additional transformations

Practical result: recommendation engines and dynamic pricing stop being "features for presentations" and start influencing conversion in real-time mode.

Why Now

The wave of AI investments in enterprise has hit the same barrier: models are good, data is bad. McKinsey and Gartner research shows that most AI pilots in retail and B2B commerce don't scale precisely because of data quality and connectivity issues. By estimates, companies spend 60–70% of AI project time on data preparation, not on model work.

SAP, which has tens of thousands of enterprise clients installed worldwide, is in a unique position: the company not only sells AI tools, but also controls the data layer on which these tools depend.

"Personalization is not an algorithmic problem, it's a data problem.

If structures are incompatible, the model won't help"—a position shared by most corporate systems architects.

What This Means

For enterprise companies on the SAP stack, a real path opens up to operational AI personalization without replacing the entire infrastructure. For competitors—Salesforce Commerce Cloud, Adobe Commerce, Shopify Plus—this is a signal: data alignment becomes a key product battleground in 2025–2026. Whoever unifies data at the execution layer first wins the contracts for AI transformation.