SAP integrates Databricks to enhance AI readiness with new Business Data Cloud

German software giant SAP is pushing the bar on the data front to power next-gen AI use cases. The company today introduced Business Data Cloud (BDC), a new SaaS product that embraces lakehouse architecture to help teams enrich their SAP ecosystem data with external data assets from different source systems and drive long-term value.

The product is the outcome of a landmark collaboration with data ecosystem major Databricks. Essentially, SAP BDC natively integrates capabilities and data from Databricks’ data intelligence platform. This removes the need for creating and maintaining complex pipelines and creates a harmonized data foundation for advanced AI agents and analytical workloads.

Several enterprises, including Henkel, are using BDC to power their AI projects. SAP itself is using the enriched BDC to power a new era of Joule agents focused on specific domains like finance, service and sales. The development makes SAP another notable player, much like Microsoft and Salesforce, bolstering its data platform to lay down the foundation for AI.

SAP’s revamped data foundation

Over the years, SAP has established itself as one of the leading players in enterprise resource planning (ERP) with S4/HANA cloud and several mission-critical applications for finance, supply chain and human capital management. These apps produce petabyte-scale data with business context and have been powering AI and analytical value for teams, via the company’s business technology platform (BTP). 

So far, SAP BTP has had a ‘datasphere’ that allows enterprises to connect data from SAP with information from non-SAP systems and eventually link it with SAP analytics cloud and other internal tools for downstream applications. Now, the company is evolving this experience into the unified BDC, natively powered by Databricks.

What SAP business data cloud has on offer

What this means is that SAP is embracing lakehouse architecture, creating a unified foundation that combines all SAP data products — from finance, spend and supply chain data in SAP S/4HANA and SAP Ariba, to learning and talent data in SAP SuccessFactors — with structured and unstructured data from other varied yet business-critical systems, stored in Databricks.

Once the data is unified (via zero-copy, bi-directional sharing), SAP BDC can leverage Databricks-specific capabilities for workloads like data warehousing, data engineering and AI, all governed by Databricks unity catalog.

“We take all of these different data products, which are provisioned and managed by SAP…and we will persist them into the lakehouse of SAP business data cloud, in a harmonized data model,” Irfan Khan, president and CPO for SAP data and analytics, told VentureBeat. “This lakehouse will have Databricks capabilities for users to build upon.”

Previously, said Khan, users who had a large percentage of their data in Databricks and SAP data in S4 or BW had to build and manage complex pipelines and replicate all the data assets to the SAP platform while rebuilding the entire semantics and the core data model at the same time. The approach took time and required them to keep their pipelines updated with changing data. However, with Databricks’ native integration, users have access to everything in one place and can directly do data engineering, data science and other tasks on top of the BDC.

“In Datasphere, you had a means of doing a similar thing, but they were all customer-managed data products,” Khan explained. “So, you had to go into the data platform, select the data sources and build the data pipelines. Then, you had to figure out what to replicate. Here, it’s all managed by SAP.”

What this means for enterprises

At its core, this Databricks-powered product gives teams a faster, simpler way to unify and mobilize their business data assets locked within SAP and Databricks environments. 

The combined, semantically-enhanced data will pave the way for building next-gen AI applications aimed at different use cases. For instance, a team could use Databricks’ Mosaic AI capabilities to develop domain-specific AI agents that could use context from SAP’s business data as well as external Databricks-specific data to automate certain human capital management or supply chain functions. 

Notably, SAP itself is tapping this enhanced data foundation to power ready-to-use Joule agents aimed at automating tasks and accelerating workflows across sales, service and finance functions. These agents deeply understand end-to-end processes and collaborate to solve complex business problems.

Beyond this, BDC will have an “insight apps” capability, which will allow users to connect their data products and AI models with external real-time data to deliver advanced analytics and planning across business functions.

More data partners to come

While the partnership underscores a big move for both Databricks and SAP, it is important to note that the Ali Ghodsi-led data major won’t be the only one bolstering BDC. 

According to Khan, data sharing and ecosystem openness are the company’s first design principles — and they will expand to other data platforms through their partner connect capabilities. This means an enterprise user will be able to choose the platform they prefer (or that they are locked into) and bi-directionally share data for targeted use cases.