Concept card

SAP Business Data Cloud data products as AI context

As of 2026-09-25

Edited by Cédric-Olivier Mâry, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is SAP Business Data Cloud data products as AI context?

A BDC data product is a governed, versioned, semantically described SAP asset — not a raw table export — and it becomes AI context through zero-copy sharing to specific consumers: Databricks (GA October 2025), Google BigQuery (public preview), Snowflake (from H1 2026), or direct exposure inside Datasphere. "AI-ready" describes the governance and semantics, not just the plumbing that moves the bytes.

What a data product is, precisely

A SAP Business Data Cloud data product is not a table dump. It is a governed, versioned, business-semantic asset — SAP's own framing, echoed across its 2025-2026 announcements, is that these products carry structure and business logic with them, not just rows and columns. That distinction is the whole point of this card: an AI agent or model that consumes a data product inherits field names, relationships and business rules that SAP maintains, instead of a data engineer having to reverse-engineer them from a raw extract.

SAP's May 2025 Sapphire announcements described hundreds of new data products launching through 2025 across Business Suite domains, including environmental, health, safety and product-compliance datasets — evidence of breadth, not a claim of complete coverage. A consultant scoping an AI use case should verify, for the specific domain in question, that a data product actually exists and is activated, rather than assuming BDC has "everything" — coverage is real but uneven, and the honest move is to check the live catalogue for the engagement at hand rather than quote a headline count that ages within weeks.

Why it matters

A data product carries SAP's own semantics and governance; treating it as a plain table export throws away exactly the guarantee that makes it worth using over a custom extract.
Zero-copy sharing is now GA to Databricks (6 October 2025) and extending to Google and Snowflake (H1 2026) — the data-movement problem is largely solved across major platforms, which shifts scoping effort toward activation and governance instead of pipeline engineering.
"Hundreds of new data products" (2025) is real breadth, not full coverage — every AI engagement should verify the specific domain's data product exists and is activated before designing around it.

Key points

A data product is governed, versioned and carries SAP's business semantics — not a raw table dump.
Zero-copy to SAP Databricks: GA 6 October 2025, built on Delta Sharing, secured with mTLS + OAuth, mounted into Unity Catalog.
Zero-copy to Google BigQuery: also part of the BDC Connect family; SAP-Snowflake: solution extension Q1 2026, BDC Connect for Snowflake H1 2026.
Inside Datasphere, a data product can be modeled and exposed to Just Ask/Joule (path 2 of C348) without ever leaving the SAP estate.
"AI-ready" is SAP's language for governance and semantics, not just data movement — zero-copy solves duplication, not ambiguity or access control.
The BDC Administrator role gates which data products in a package are activated and visible in the catalog.
Hundreds of new data products launched through 2025 across Business Suite domains — verify current coverage per engagement, do not assume completeness.

Terms used on this page

Data product: A governed, versioned, business-semantic SAP data asset activated and shared from Business Data Cloud.
Zero-copy sharing: Live data access at the source without replication, built on Delta Sharing across BDC's Databricks, Google and Snowflake connects.
BDC Administrator: The role that activates which data products in a package are made available in the catalog.
Delta Sharing: The open sharing protocol underlying BDC's zero-copy connects to Databricks and other platforms.
Business Data Fabric: SAP's strategic framing for an open, AI-ready data ecosystem spanning SAP and third-party platforms.

Sources

Full card available to members. What the full card adds: the full decision framework · the common pitfalls and their fix · the cheat sheet · the facts worth quoting.

Open in the app →