SAP data cloud, Databricks integration aims to unify AI data
SAP unveiled the Business Data Cloud platform and Databricks partnership to support customers in AI projects, which analysts bill as an important evolution of data management.
SAP launched a new offering and partnership with Databricks aimed at enabling companies to prepare enterprise data for AI capabilities.
The new SAP Business Data Cloud (BDC), billed as an evolution of the SAP Datasphere data fabric platform, is a fully managed SaaS designed to unify the governance and structure of SAP data and connect to external data to provide more comprehensive data for AI applications.
Databricks' data management technology is embedded into BDC to help bring external data into the platform. The new offering and partnership were introduced Thursday at SAP Business Unleashed, a virtual event.
The new data platform and partnership with Databricks are intended to help customers deal with the challenges they have faced in implementing and getting value from AI technology, according to Muhammad Alam, head of product engineering at SAP, in a media briefing.
"It's hard to get reliable, high-quality data sets across the enterprise to power this AI," Alam said. "Many have done hundreds of proofs-of-concepts but have not been able to get them into production, and the ones that have been able to get into production haven't seen the adoption that they're looking for to be able to realize the value."
SAP BDC includes prebuilt data products that span across the SAP suite of applications, including finance, procurement and supply chain data in S/4HANA and SAP Ariba, along with human capital management data in SAP SuccessFactors.
The data products are comprised of data from SAP applications that maintain the original business context and semantics of the data that customers can use out-of-the-box, Alam said.
Databricks introduces external data
The Databricks partnership enables customers to combine SAP data with external data by embedding Databricks capabilities into BDC, according to Alam.
"With [SAP and Databricks], you get the world's most important data set with the world's leading data platform capabilities," he said.
The goal is to enable SAP customers to run applications on combined data sets for data science, AI, machine learning and SQL, according to Shanku Niyogi, vice president of product management at Databricks.
"That means that they can get the benefit from unified data AI governance, build intelligent applications using the Mosaic AI stack or use Databricks Notebooks to combine data from different sources and create derivative data products," Niyogi said.
Customers using Databricks will be able to more easily access data from SAP and combine it with their lakehouse data to do analysis and build data products or AI applications, he said.
SAP also introduced a series of Joule generative AI agents for tasks across SAP business applications such as processing claims, accelerating dispute resolution and answering customer service calls, according to Alam.
In addition, SAP is previewing an agent builder capability that enables customers to extend Joule agents or build their own. The agent builder uses SAP Knowledge Graph, a data management capability in SAP Datasphere that maps and correlates the relationships and context across the systems, he said.
An important SAP development
SAP BDC is the most important platform innovation for SAP in the last 20 years, said Holger Mueller, vice president and principal analyst at Constellation Research.
The SAP HANA in-memory database was seen as an important innovation, but HANA was developed as a replacement for underlying databases of SAP ERP systems and did not expand SAP's ability to build the business applications of the 21st century, Mueller said.
"With SAP BDC, SAP goes back to partnering for a platform -- in this case with Databricks -- and it is more than relevant for SAP customers as it allows them to address the non-structured data challenges," he said.
SAP BDC has created the data foundation for SAP to build its next generation ERP applications and it may create the incentive and the necessity for SAP customers to migrate to the cloud, Mueller said.
"SAP BDC is the long awaited, more than necessary SAP data extension, giving customers the option to use SAP offerings for this instead of building lakehouses on their own," he said. "It's good for customers to have options."
SAP BDC will need to prove its value for customers, but SAP is fundamentally correct in its contention that effective enterprise AI requires a coherent data platform, according to Jon Reed, co-founder of Diginomica, an enterprise industry analysis firm.
"You can't have AI scarfing up data and giving you outputs that have role-based parameters, you need to have all of that in there," Reed said. "The more you have a unified data platform with all of that information, the better off you are."
SAP is evolving how it can better serve customers with data for AI, which means thinking of itself not just as an applications provider, but as a data provider that takes burden off of customers and differentiates itself from general-purpose AI providers, he said.
"Why couldn't a third-party LLM provider like Cohere come into SAP's customer base, set up an LLM and start feeding all the data," Reed said. "That's what SAP's up against."
Still, customers will have the final say on whether to adopt the new offering, and they will take everything into account, including the accuracy of the results, the overall value and the cost, he said.
"It's good that SAP is trying to take responsibility and ownership of these data challenges for customers, but we still have to see how well they'll do with that," Reed said.
Jim O'Donnell is a senior news writer for Informa TechTarget who covers ERP and other enterprise applications.
