Most companies are still in the experimental phase with AI. Many of them already have impressive demos, working pilots and tools that help them complete specific tasks even faster. Significantly fewer companies can say that AI has led to operational changes across all areas, processes and teams.
The decisive factor for the successful use of AI in a company is not the model, but the context: the ability of AI to understand business processes.
When people talk about AI today, it's mostly about agents, models and benchmarks. Which model achieves the best results? Which system performs the most tasks? Which interface is the most intuitive? While these factors are important, they do not solve the core challenges facing companies.
Organizations run complex workflows that span multiple teams, policies, approvals, permissions, and data. They use systems that are central to their business activities for their planning, procurement, production, hiring, payments and services. AI only creates lasting added value on a broad basis if it operates within this structure.
Models generate answers. An agent can perform a task. But that is not enough to control a company's business processes. This requires understanding how tasks are carried out, who is authorized to take actions, what rules apply, and how decisions interact across different areas. Without this context, AI cannot live up to expectations.
This is just one reason why I am convinced that the use of AI will make software with comprehensive business context even more important. It enables companies to completely redefine work processes. When AI agents understand end-to-end processes, they can work across departments, execute workflows independently and coordinate actions in real time. Instead of just automating individual steps, AI can manage processes end-to-end. This relieves employees of repetitive coordination tasks and allows them to concentrate on important decisions, supervisory activities and strategies.
We refer to this as that autonomous companiesa fundamental shift from classic enterprise systems to systems that can draw conclusions, make decisions and act. A vision in which SAP plays a pioneering role.
We have been mapping the core processes of the world's leading companies for more than five decades. Our systems don't just store data. It also stipulates how companies work: their processes, rules and decisions. Our ERP systems are the institutional memory and brains of many companies in all industries around the world. Our new one SAP Business AI Platform unifies this corporate data, processes and governance in a unified context for AI.
On this basis, Joule is the interaction layer that connects people with AI and enables them to work with software in completely new ways. Joule assistants work with users, while Joule agents execute business processes end-to-end. In this way, intelligence is embedded directly into the processes and not just added on. We call this the SAP Autonomous Suite.
“Show me how my financial forecast for the year could change based on the latest sales pipeline and supply chain data.At first glance, this looks like a simple prompt for a large language model, but without connection to enterprise systems, the answer is pure speculation.
Given the complete business context, the system first identifies the correct business process from hundreds of business-critical processes and understands the specific configuration of how that process is executed in the business. It then selects exactly the right data from millions of data fields stored across the ERP landscape. Finally, identity, authorization and access are checked at every step and ensure that the result is correct, compliant and reliable. In this way, companies not only receive general answers based on probabilities, but also make decisions they can trust.
To achieve this state, it is not enough to simply add a chatbot or subsequently integrate AI into existing systems. Many organizations still operate with fragmented landscapes, data spread across many systems, and processes characterized by years of incremental change. In this environment, AI does not ensure faster progress and can actually increase inefficiency and risk. Companies must redesign the interaction between processes, data and infrastructure and clearly define the responsibility between people and AI. This is not just a technical change. It is a change management task.
New technology only creates added value if it is accompanied by real change. AI is not a replacement for transformation. It increases the benefits of a successful transformation. The benefits are only fully realized when all elements of the system – the agent, the process and the person – work together seamlessly. Users must understand how to work with AI agents, and processes must be intentionally designed to integrate intelligence where decisions are made and tasks are carried out.
This is why change management is fundamental. It includes upskilling the workforce, redesigning processes to link them directly to data and AI, and modernizing the underlying system landscape.
For this reason, we are introducing new AI-supported offers via RISE with SAP and SAP GROW and adapt our service model: to help companies modernize, master change and create sustainable value through AI at their own pace.
This is the beginning of a new era of enterprise software in which intelligence is not separated from processes, but embedded in them. It will not be the companies with the most advanced models that will set the tone here, but rather those that link AI to their actual business processes – with context and clear governance.
This is the dawn of the autonomous enterprise and SAP is well positioned to pave the way there.
Christian Klein is CEO of SAP SE.
