SAP: AI Without Data Context, a Risk for Businesses

The Rise of AI in Businesses

Artificial intelligence (AI) is experiencing rapid adoption in the business sector, transitioning from a phase of experimentation to everyday use. Organizations are now deploying copilots, agents, and predictive systems across various fields such as finance, supply chains, human resources, and customer operations. According to a recent survey, by the end of 2025, half of businesses are expected to use AI in at least three business functions.

The Challenge of Data Context

As AI increasingly integrates into critical business processes, leaders find that the main obstacle is not the performance of models or computing power, but rather the quality and context of the data on which these systems rely. AI introduces a new requirement: systems must not only access data but also understand the business context surrounding it. Irfan Khan, President and Chief Product Officer of SAP Data & Analytics, explains that without this context, AI can generate responses quickly but make poor decisions.

The Need for Context

"AI is incredibly effective at producing results," says Khan. "It moves quickly, but without context, it cannot exercise good judgment, and good judgment is what generates a return on investment for the business. Speed without judgment is not useful. It can even harm us." In this new era of autonomous systems and intelligent applications, this layer of context becomes essential. To provide this context, businesses need a well-designed data architecture that does more than just integrate data. The right data architecture allows organizations to scale AI safely, coordinate decisions between systems and agents, and ensure that automation reflects true business priorities rather than making isolated decisions.

The Loss of Context: A Critical Problem for AI

Traditional data strategies have largely focused on aggregation. Over the past two decades, organizations have invested heavily to extract insights from operational systems and load them into centralized warehouses, lakes, and dashboards. This approach facilitates reporting, performance tracking, and generating insights across the enterprise, but in the process, much of the meaning attached to this data—how it relates to policies, processes, and real decisions—is lost.

Consider two companies using AI to manage supply chain disruptions. If one uses raw signals such as stock levels, delivery times, and supply scores, while the other adds context through business processes, policies, and metadata, both systems will quickly analyze the data but likely arrive at different conclusions. Insights such as which customers are strategic accounts, what trade-offs are acceptable during shortages, and the status of extended supply chains will enable one AI system to make strategic decisions, while the other will lack the appropriate context, explains Khan.

Preparing for AI

This awareness is changing how businesses view their readiness for AI. Most recognize that they do not have the mature data processes and infrastructure necessary to trust their data and AI systems. One in five organizations considers its data approach to be very mature, and only 9% feel fully prepared to integrate and interoperate with their data systems.

Towards a Data Architecture

The emerging solution is a data architecture: an abstraction layer that spans infrastructure, architecture, and logical organization. For agentic AI, this architecture becomes the main interface, allowing agents to interact with business knowledge rather than raw storage systems. Knowledge graphs play a central role, enabling agents to query enterprise data using natural language and business logic.

The value of this architecture rests on three components:

• Intelligent computing to provide speed
• Knowledge reservoir to provide business understanding and context
• Agents to provide autonomous action grounded in this understanding

What makes this powerful is how these capabilities interact, according to Khan.

Conclusion

Technology provides the architecture—a foundation that enables communication and coordination among agents. The process will define how businesses and IT share ownership, establish governance, and cultivate a culture of trust that encourages adoption. All these elements must now work together for an enterprise data architecture to be truly successful. "It enables confident and consistent decisions, and when these elements come together, AI does not just analyze and interpret data—it makes smarter and faster decisions that truly create business impact," he concludes.