Updated
Updated · MIT Technology Review · Nov 16
SAP, Infosys Push Enterprise AI Beyond Pilots as Only 14% of Firms Scale Successfully
Updated
Updated · MIT Technology Review · Nov 16

SAP, Infosys Push Enterprise AI Beyond Pilots as Only 14% of Firms Scale Successfully

3 articles · Updated · MIT Technology Review · Nov 16

Summary

  • Only 14% of large enterprises have successfully scaled AI, SAP and Infosys said, pitching a joint approach to move companies from fragmented pilots into production AI embedded in core workflows.
  • 74% of organizations are still stuck in pilot mode, they said, largely because AI projects start as isolated experiments without a unified data layer, business-process context, or governance.
  • SAP is positioning its Business AI Platform as that foundation, combining data integration, access to 50-plus models, and agent governance so companies can manage AI tools from build to retirement.
  • Infosys said its preconfigured use cases and implementation work can cut deployment times to 4-6 months from 12-18 months, citing examples including a 92% reduction in SAP specification work.
  • The partnership argues the next phase is AI as a decision-making partner, with governed agents acting as digital coworkers across finance, supply chains, and other mission-critical enterprise functions.

Insights

If autonomous AI agents require perfect enterprise data to function, are most legacy companies doomed to fail before they even start?
Could the heavy centralized governance required for agentic AI actually stifle the rapid innovation that made generative AI so appealing?
When a digital coworker makes a costly mistake, who is ultimately held responsible in a fully automated enterprise workflow?