Enterprise AI Failure Rates Rise to 42 Percent

Enterprise AI initiatives are experiencing a sharp increase in failure rates, rising from 17 percent in 2024 to 42 percent in 2025, according to research from S&P Global. This trend suggests a growing gap between the technical capabilities of AI models and the ability of organizations to implement them at scale.

Data from EXL Service indicates that approximately 60 percent of C-suite executives across the banking, insurance, retail, utilities, and healthcare sectors report that their AI projects remain stuck in the pilot phase. These failures are not attributed to technical flaws in the AI models themselves, but rather to an inability to integrate the technology into existing workflows.

Primary obstacles include siloed data, incompatible legacy systems, and a lack of specialized talent. To address these issues, industry leaders suggest a strategy focused on establishing clean data foundations, using domain expertise to reengineer business processes, and developing AI-driven operating models that reimagine entire workflows instead of deploying isolated point solutions.