The end of data silos? How SAP is redefining enterprise AI with Joule and Databricks

SAP today announced a substantial upgrade to its enterprise-wide data management offerings and strategy. Called SAP Business Data Cloud, the new offering is a managed software-as-a-service (SaaS) that, the company says, “unifies and governs” all SAP data, “seamlessly” connects it with third-party data, and provides AI capabilities throughout.

The company also announced a strategic partnership with Databricks, which creates unified data lakes and data warehouses (they call them Lakehouses). Databricks’ data unification makes it possible for more comprehensive AI workloads and analytics.

Also: Crawl, then walk, before you run with AI agents, experts recommend

In a separate but concurrent announcement, SAP introduced ready-to-use Joule agents for service, sales, and finance. Joule is SAP’s conversational AI assistant.

Make no mistake: These are big announcements, especially for existing SAP customers. Let’s break down what this all means in the context of the SAP ecosystem.

Going beyond SAP Datasphere

To be clear, SAP has long offered a data unification and management solution in its SAP Datasphere offering. SAP Datasphere is marketed as a business data fabric that allows real-time connectivity between systems and applications.

According to SAP’s Datasphere product page, “SAP Datasphere capabilities are natively available in SAP Business Data Cloud.”

SAP Datasphere provides data management and integration across both SAP and external systems, while the new SAP Business Data Cloud is a much broader, fully-managed SaaS platform that not only manages SAP data but incorporates AI and deeper analytics integration using the data lake and data warehousing capabilities of Databricks.

In terms of strategic direction, SAP Datasphere focuses on connectivity, while the new SAP Business Data Cloud platform provides advanced capabilities for governance, wide interoperability, and AI-driven insights and oversight. Key to the SAP Business Data Cloud is the Databricks partnership, which enables SAP clients to operate a much more seamless, AI-driven data environment.

The Databricks partnership

Pop quiz: When it comes to enterprise insight and governance, what is AI’s kryptonite? Short answer: data silos.

The thing that has made ChatGPT so amazing is its ability to sift rapidly through what seems like the entire web and synthesize answers based on a broad and ever-growing knowledge base.

When you’re managing a sprawling enterprise, you want to be able to get answers and insights that are as broad and comprehensive. But if much of your data is stored in individual silos — operated by different business units, departments, and even third parties using unrelated solutions — all that data becomes invisible to comprehensive AI analysis.

Also: AI data centers are becoming ‘mind-blowingly large’

That’s where the Databricks partnership comes in. Databricks’ core competence is toppling silos to provide a more universal view of both SAP and non-SAP data. The proprietary Lakehouse architecture is built on top of Apache Spark’s distributed data processing engine, Delta Lake’s reliable and performant storage layer, and MLflow, the machine learning lifecycle management tool, among other technologies.

By combining Databrick’s flexible data sharing with SAP’s strong governance capabilities, data governance — especially compliance and security — is simplified and reinforced.

Databricks also offers far deeper business insights, both because AI is no longer stymied by data silos, and because Databricks has its own AI and large-scale analytics capabilities that are built for deeper insights and predictive decision-making across enormous oceans of data.

Essentially, the Databricks partnership gives SAP’s AI room to run. That’s critically important to enable.

Joule agents

SAP is introducing pre-built AI agents that actively perform tasks rather than simply providing recommendations. They are designed to automate and execute complex business processes.

Before we go on, I’ll share with you my concern about AI agents. We know AIs make mistakes and confabulate. When working with a chatbot, you have to make sure you double-check every so-called fact the AI provides.

Now, what happens when you scale that up and let an AI loose across your enterprise? There’s always the possibility the AI will make a serious mistake and then propagate that mistake at light speed all across your network.

Also: AI agents might be the new workforce, but they still need a manager

To be fair, humans are equally prone to mistakes. Here at ZDNET we have written regularly about how human error caused services and networks to go down until repaired. Just don’t start deploying agents thinking they’ll be perfect. Be sure to QA everything, human and AI.

OK, now that we’ve passed my “trust, but verify” warning for AIs, let’s discuss Joule agents.

SAP is introducing pre-built agents for the finance, sales, and service industries.

The finance agent can help handle repetitive financial tasks and help in decision-making. SAP specifically called out a cash collection agent, which can analyze disputes and “work across finance, customer service, and operations to validate details and recommend resolutions.”

In the future, finance agents might be able to perform automatic invoice processing, matching invoices to purchase orders, identifying anomalies, and pointing out discrepancies. They could also provide more predictive cash flow management, help with fraud detection, automate expense management, and reconcile accounts to help with the monthly, quarterly, and yearly closing processes.

Also: AI agents will match ‘good mid-level’ engineers this year, says Mark Zuckerberg

The Joule service agent works with customer service organizations to resolve customer service issues. Using contextual awareness from the SAP Knowledge Graph, the AI can analyze previous service interactions to suggest relevant solutions.

Going forward, agentic AI could help with automated case resolution, proactively prevent equipment failures by suggesting maintenance procedures predictively, help customers provide self-service solutions, optimize schedules for service technicians and replacement parts, and detect rising customer dissatisfaction by identifying the root problems causing that dissatisfaction and suggesting mediations.

SAP is introducing a cross-functional Q&A agent that can monitor sales opportunities and customer cases, proactively identifying questions and providing answers from appropriate knowledge sources. This agent can support both the sales and service teams.

Another agent is a case creation agent that actively watches customer service cases; when a new way of resolving a problem is identified, it creates a knowledge article that customer service staff can later reference.

Also: 7 apps that helped me escape the cloud – and protect my data privacy

A case classification agent can understand the context of a case (SAP uses the example of identifying a tax-related query even if the word “tax” isn’t used) and then route the cases to the most qualified teams or individuals.

Assuming appropriate guardrails are in place, AI agents, particularly those vertically integrated with deep and validated solutions like SAP offers, have tremendous potential to reduce cost and time while increasing output quality.

Four key business data takeaways

There seem to be four major strategic legs to SAP’s modernized data strategy. They are:

• Business semantics: This strategy basically goes to the idea of all the data being accessible, but kept in place and in context. In other words, rather than sucking out and normalizing tons of data, without being able to find their origin records, the new systems can model, share, and model data without extraction. This is called the “zero-copy” approach.
• Data engineering with AI: AI models can be optimized by SAP integration with Databricks’ managed lakehouses using a technology called Delta Sharing. According to SAP, the approach “harmonizes SAP data products with existing lakehouses bi-directionally.” This gets over the age-old problem of migrating data out of one system to work on it, but then not being able to keep the original system updated or relevant.
• Warehouse modernization: For on-premise SAP Business Warehouse customers, the SAP Business Data Cloud solution can create a hybrid approach, where data can still be stored in the on-prem warehouse, but accessed as a data product within an object store using Delta Share. This enables customers to continue to derive value from the original on-prem data warehouse investment while gaining data lake benefits.
• Analytics and planning: The new solution brings together analytics, reporting, and AI-driven forecasting, supporting real-time financial, supply chain, and operational planning, all from one platform.

Integration with partner apps

SAP describes Business Data Cloud as “built to prioritize openness and customer choice” as an open data ecosystem. At announcement time, the environment integrates natively with solutions from Collibra, Confluent, and DataRobot, as well as solutions provided by McKinsey, PwC, EY, Deloitte, and Capgemini.

What this means for your enterprise

SAP Business Cloud and Joule AI agents are offerings that will help you make a more strategic shift toward an intelligent, interconnected enterprise powered by AI-driven data analytics, driving a holistic decision-making process enterprise-wide. SAP is making a big push to unify structured and unstructured data with advanced AI agents and analytic operations across systems and even providers.

Also: How ChatGPT’s data analysis tool yields actionable business insights with no programming

SAP is not simply enhancing business efficiency, but redefining how their customers extract value from data and scale AI-driven operations in our increasingly complex digital landscape.

What do you think about SAP’s latest moves? Do you see SAP Business Data Cloud and the Databricks partnership as a game-changer for enterprise AI and data management? Are you excited or skeptical about Joule AI agents handling critical business tasks? How do you think AI-driven automation will impact finance, sales, and customer service in the years ahead? Let us know in the comments below.

You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.