Christiane Grimm
In the SAP world, everything went according to plan for a long time: There were clear data structures, fixed processes, and classic reporting tools. But today it is clear that this is no longer enough. With the SAP Business Data Cloud and the integration of Databricks, a new data universe is opening up that connects the structured SAP world with the flexible AI galaxy. In an expert interview with Sven Knöpfler, you can find out why now is the right time to rethink data, AI, and analytics.
Sven, you've been working in consulting for over twenty years now, and during that time you've been involved in countless SAP projects – from classic on-premise implementations to the transformation to modern cloud platforms. So you've experienced firsthand how the SAP Analytics world has changed. Now, suddenly, a new player is emerging in this cosmos: Databricks.
Before we dive deeper, could you give us a quick overview and explain: What is Databricks?
Gladly! Let's imagine Databricks as the "brain of the company". A brain can perceive, understand and anticipate. Transferred to companies, this means:
- Perceiving: This means recording all relevant signals, i.e., data. This no longer only involves the clean, structured data from the SAP ERP system, but also unstructured information such as emails from customers, images, social media feeds, or sensor data from machines.
- Understanding: The brain processes these signals and puts them into context. This is where artificial intelligence comes into play. Using classic AI and statistics, we can recognize patterns in structured data. The new multimodal Generative AI, on the other hand, can understand and link the content of texts, images, and speech simultaneously.
- Anticipate: This is the highest level. On this basis, the brain can make predictions or give recommendations for action. For a company, this means acting proactively, for example through precise sales forecasts, recommendations for the next marketing campaign, or predictive maintenance of a machine before it fails.
You just said Databricks is the 'brain of the company' – powerful, flexible, and AI-driven. What specifically makes it different from classic SAP technologies like BW or SAC?
Databricks is a leading, unified data and AI platform based on a so-called lakehouse architecture. This term is crucial: A lakehouse combines the strengths of data warehouses (like SAP BW – high data quality, structured analyses) with the advantages of data lakes (huge amounts of any data, flexibility, cheaper storage).
What makes it so special compared to classic SAP technologies like SAP BW? Here, the brain analogy comes into play again:
Perception (Data Acquisition): Classic SAP technologies excel at processing structured data from SAP sources. However, they are cumbersome when it comes to unstructured data like documents, images, or IoT streams. Databricks is built precisely for this purpose. It can effortlessly ingest and process large volumes of structured and unstructured data, dramatically expanding your company's 'senses'.
Understand (Analysis & AI): While classic SAP technologies are primarily designed for descriptive analysis and reporting (What happened?), Databricks is a fully-fledged AI workshop. It enables not only classic machine learning but is also a leader in the field of Generative AI. Here, you can train models that understand the content of a damage report, analyze a product image, and link this to the order data from SAP. This deep, multimodal understanding was previously not possible in the pure SAP world.
Anticipate (Predict): Because Databricks has all the data in one place and can understand it with state-of-the-art AI, the predictions and recommendations it generates are much more powerful. It's no longer just about a simple trend extrapolation of sales figures. It's about simulating complex scenarios and providing proactive recommendations – a core capability for an intelligent company.
Impressive! – especially with regard to AI and data diversity.
But how exactly does it fit into the SAP world?
Is it more of an addition to the BDC, a technical integration – or are we even talking about a completely new way of handling data here?
The strategic partnership between SAP and Databricks has resulted in «SAP Databricks»: an OEM solution that is an integral part of the SAP Business Data Cloud.
SAP Databricks fulfills several roles:
- Addition: SAP Databricks complements the strengths of the BDC. The BDC remains the leading system for providing SAP data with all its semantics for analytical purposes. SAP Databricks adds big data and AI capabilities that SAP did not natively possess in this depth before.
- Integration: The partnership ensures deep technical integration. Data no longer needs to be copied in complex, overnight processes. With SAP Databricks, you can directly and efficiently access SAP data and combine it with non-SAP data. This speeds up projects from months to weeks.
- New way of thinking: This is the most important point. We solve business problems by using the best data, not just the most easily accessible. Only those who know the whole picture can make informed and forward-looking decisions.
A new way of thinking, deeper integration, strategic partnership – you can sense that something fundamental is changing.
For a long time, SAP data was almost exclusively used within its own environment – now, new doors are suddenly opening. Why is now the right moment to consider SAP data outside of this familiar landscape, for example, in a platform like Databricks?
The timing is not accidental, but the result of several megatrends coming together:
- The AI revolution: Above all, Generative AI has fundamentally changed expectations of what is possible with data. Companies are realizing that there is enormous added value in connecting valuable SAP process data with the outside world (documents, images, market data). SAP Databricks is the platform that makes exactly that possible.
- Data explosion: The sheer amount and variety of data (IoT, social media, etc.) overwhelms traditional architectures. An SAP BW is not built for processing terabytes of sensor data in real time. The pressure to use all this data is increasing.
- Urgent need for agility: The combination of SAP and Databricks makes it possible to implement new applications in a very short time and react to changes. The cloud architecture makes it possible to scale quickly - up and down.
- Strategic Opening of SAP: Instead of trying to build everything themselves, SAP relies on strategic partnerships with the “Best-of-Breed” providers in selected areas. The partnership with Databricks is not the only, but certainly the most prominent example of this opening strategy.
You've described how SAP and Databricks strategically complement each other and the thinking behind it. That sounds promising – but what does that mean in the practical day-to-day of IT and data architecture?
What becomes technically possible all of a sudden, which was hardly conceivable before in typical SAP architectures?
Technically, things are becoming possible that were previously extremely complex or simply impossible:
Advanced Analytics on Real-Time Data: Imagine being able to run an AI analysis to predict supply chain bottlenecks directly on the logistics data in your S/4HANA system, and enrich it in real time with external data such as weather forecasts, ship trackers, or news feeds. Previously, you would have had to painstakingly extract and consolidate data, which always resulted in a time delay. Today, it works in real time on an integrated platform.
Agentic AI Use Cases: A customer sends an email complaint with a photo of a damaged product. A GenAI application in Databricks can analyze the text of the email (sentiment, problem description) and the image (type of damage), compare it with the master data of the customer and the product from SAP, and automatically create a service order in SAP, suggest the right technician, and reserve the appropriate spare part. This is end-to-end automation that overcomes silos.
Scalable End-to-End Machine Learning Operations: Previously, the development of AI models in the SAP environment was a patchwork of different tools for data extraction, model training, and commissioning. With SAP Databricks in the SAP BDC, you have an integrated platform to manage the entire lifecycle of an AI model – from data preparation to training and productive monitoring. This brings a level of professionalism and speed to the process that was previously not possible.
This shows how profoundly the technological basis has changed. Now, looking at it personally: Is there a particular aspect or use case that particularly excites you – perhaps even a little “Nerd-Moment”?
What personally excites me the most are the possibilities that will arise from the combination of GenAI, Agentic AI, and the SAP Datasphere Knowledge Graph. This will intelligently connect all corporate knowledge – structured SAP data and unstructured information such as images or texts – and equip it with business context.
Imagine an AI agent that not only recognizes problems but immediately understands the entire operational impact. It autonomously informs all affected stakeholders and acts in coordination with humans to solve the challenges end-to-end.
The historical task of IT has always been to automate processes. For me, this is the next logical step in evolution. And analytical data architecture plays a central role in this.
It's clear that the possibilities are vast, but that's precisely what makes it challenging to take the first step. Especially for those deeply rooted in the SAP world, the question arises of how to approach this change in a meaningful way.
What would you advise someone who is working in the SAP world and notices: "Something is happening with data right now - but I don't know where to start"?
My advice consists of three simple steps:
Start with the business problem, not the technology. Ask: “What is our most pressing business problem that we cannot solve with our current analytical tools?” Is it high customer churn? Is it inaccurate sales forecasts? Is it inefficient warehousing?
Think big, but start small. Start with a small, manageable use case. For example, enrich your SAP sales data with external market data to improve the forecast for a single product. This way, you quickly generate a visible success, get to know the technology and create acceptance within the company.
Build bridges. Bring the SAP experts, the specialist departments and the data scientists together at one table. The real magic happens not in the technological silos, but at the interface between process knowledge and data knowledge. Invest in the further training of your employees so that they understand the new possibilities and can ask the right questions.
Anyone who has become curious now may be wondering: “How can I delve deeper into this?” Fortunately, there will soon be just the right opportunity to do so – and you are even a speaker at our Analytics Online Conference 2025! In conclusion: What awaits us at the Analytics Online Conference 2025 – and why should you definitely not miss it?
This conference is the perfect place for anyone at the crossroads between the traditional SAP world and the new data and AI world. The integration of SAP and modern data platforms like Databricks is not a temporary trend. Hybrid architectures are the future of enterprise IT. This conference is the best opportunity to get up to speed in a short amount of time, ask the right questions, and network with like-minded people and experts. It is an investment in the future viability of your company.
