Enterprise Generative AI Investments Fail to Deliver P&L Impact

Massachusetts Institute of Technology Project NANDA reports that enterprise generative AI investments, estimated between $30 billion and $40 billion, are largely failing to deliver measurable P&L impact. According to the State of AI in Business 2025 report, 95% of generative AI pilots have shown no result.

This trend suggests a broader instability in the sector. Gartner predicts that over 40% of agentic AI projects will be canceled by 2027, citing escalating costs, inadequate risk controls, and unclear business value. These failures are particularly acute as companies transition toward agentic AI capable of autonomous action.

Su Belagodu, creator of the HITL Maturity Model, attributes these failures to the "warm body problem." This occurs when human oversight is maintained for compliance purposes but lacks actual judgment because reviewers suffer from fatigue and over-trust. Belagodu argues that surviving projects will be those that treat human oversight as a measurable production system with specific KPIs rather than a documentation exercise. This approach aligns with the National Institute of Standards and Technology AI Risk Management Framework and Article 14 of the EU AI Act, both of which emphasize the necessity of effective, managed human oversight.