Enterprise AI adoption is accelerating, but scaling agentic AI remains elusive. While 88% of companies now use AI in at least one business function, only one in 10 have successfully scaled their AI agents, according to McKinsey's latest AI report. The bottleneck isn't model quality but data infrastructure that lacks business context.

Experts warn that most companies face delays not from AI model limitations but from missing data architectures. Irfan Khan of SAP Data & Analytics emphasizes that AI agents require a modern data foundation delivering both data and business context. Without this semantic layer, agents cannot reliably execute tasks across supply chains, financial planning, and other critical operations.

The solution involves building a business-aware data fabric that harmonizes information across multiple sources. Legacy architectures cannot support autonomous AI systems, with only four in 10 companies believing their data management is AI-ready. Success requires prioritizing governance, semantics, and openness over vendor lock-in, while focusing on less-critical processes for early wins.