SAP is expanding its Business Data Cloud with tools to query data across systems without moving it, alongside new ways to give AI agents a more useful picture of how a business works. The changes matter because agents need more than access to data: they need its relationships, rules and context too.
SAP Watch analysis
What happened
In an announcement from SAP TechEd Berlin, SAP says its Business Data Cloud now supports Apache Iceberg and a universal catalogue powered by Apache Polaris. With Dremio’s SQL engine, users can query data across lakehouses and databases including SQL Server, MongoDB, Oracle and Postgres without moving it. SAP also says Dremio’s Reflection technology can automatically optimise queries as workloads change.
Other changes include expanded integration with Databricks, bringing in tools such as Lakeflow, Agent Bricks, Unity Gateway and Lakebase. SAP says its completed Reltio acquisition adds multi-domain master data management to the platform, while a new Ontology Manager lets teams create domain-specific views of SAP Knowledge Graph for agents and other applications.
Our top picks
- Query data where it lives
SAP says users can query supported databases and lakehouses without copying their data into Business Data Cloud. - Automatic query optimisation
Dremio’s Reflections are designed to learn from usage and precompute results, reducing manual performance tuning. - More Databricks capabilities
The expanded integration brings tools including Lakeflow and Agent Bricks into SAP Business Data Cloud. - Unified business records
SAP Reltio is now generally available in the platform, with AI-based entity resolution intended to match records across systems. - Context tailored to a task
The beta Ontology Manager lets teams create focused views of business entities and relationships, such as the information a collections agent needs.
Why it matters
The practical pitch is fewer data copies and more grounded answers from agents working across enterprise systems. A collections agent, for example, could be given a curated view linking customers, invoices, payment terms and credit risk, rather than being asked to make sense of scattered records on its own.
That could make existing business data more usable for analytics and AI, while keeping its context and governance in view. The capabilities and performance claims come from SAP; the announcement does not establish how they will perform across different customer environments.
Our read
This is a substantial data-platform update, not merely another promise that agents will become more autonomous. The useful idea is connecting distributed data with the business meaning needed to act on it. For buyers, the test is whether those connections work reliably with their actual systems, policies and workloads. A tidy diagram is easy; untangling enterprise data is where the plot thickens.
What to watch
- Which databases and lakehouses are supported in practice, and under what conditions.
- Whether customers report measurable gains from zero-copy queries and automatic optimisation.
- How the new Databricks integrations and SAP Reltio perform in customer deployments.
- When the beta Ontology Manager becomes generally available.
Discussion spark: For enterprise AI agents, which matters more: access to data wherever it lives, or a carefully governed map of what that data means?
Sources and evidence
- Advancing SAP Lakehouse and New Data Fabric Innovations in SAP Business Data Cloud for the Autonomous Enterprise (29 October 2026, 11:00 UTC)
not affiliated with or endorsed by SAP