Many companies have been collecting data for years but aren’t ready for the AI era. We speak with Irfhan Khan at TechEd 2025.
Those with AI plans must first lay foundations in the form of structured data. This was the common thread from Irfhan Khan, President & Chief Product Officer, SAP Data & Analytics, during a conversation at SAP TechEd. According to him, it’s “almost impossible for large companies to achieve real success with AI without a structured and formal data strategy.”
Filling data lakes
Khan outlines the situation. Many companies have built a data lake in recent years, “an empty shell” that needs to be filled with data from dozens of systems. Filling that shell requires a lot of work. And even when the data is finally there, familiar problems persist: security, governance, and quality.
When you copy data from different source systems, you lose control. Then you have to shield, anonymize, or re-verify afterward. Not ideal when you want to develop new AI applications quickly.
read also
SAP: “AI Only Becomes Truly Valuable with Context”
AI with context
What AI really needs, according to Khan, is context. He gives a great example: ask two people about a “pipeline” and you’ll get two completely different answers. For sales, it’s about deals happening, and for marketing, it’s about how leads move through the funnel. That’s why he emphasizes: “When you ask a question in the right context, you get a much better answer.”
This is precisely why SAP focuses on close collaboration between applications, data, and AI. Applications provide useful business data, the Business Data Cloud (BDC) stores it, and AI in turn uses it. “This way, data remains not just a collection of fields, but a story that the user recognizes,” says Khan.
Zero copy
One of the major steps is SAP’s zero-copy vision. Khan describes it as “a powerful architecture that eliminates the need to constantly copy data from point A to point B.” With BDC Connect, data is shared with platforms like Databricks and Snowflake without the need for repeated copying. This means companies can extract value from data much faster.
read also
SAP: “the Era of Self-Reinvention Has Arrived”
Legislation made easy
Europe thinks strictly about data. This brings necessary challenges for international companies. “You need to know exactly who has access, how data moves, what gets logged, and how everything is controlled.” SAP tries to make this process easier for companies by including access, rights, and data products as standard in the BDC architecture. This keeps companies compliant without halting their innovation.
SAP makes it clear that AI only delivers value when the underlying data is correct, available, and in the right context. Organizations that now structurally organize their data strategy will benefit from faster innovation, simpler compliance, and AI applications that directly connect to their business processes. This shifts data from loose collections to a coherent foundation on which companies can reliably deploy AI.