MIT Study Finds 95% of AI Projects Deliver No Business Return Despite Massive Investment

A recent study by the Massachusetts Institute of Technology (MIT) has revealed that despite the global rush to adopt artificial intelligence (AI) systems, the majority of these projects have failed to deliver tangible business returns. The study reported that 95 percent of organisations implementing AI solutions are receiving zero return on their investments, signaling widespread challenges in realising financial benefits from AI.

Titled “The GenAI Divide: State of AI in Business 2025,” the report surveyed 300 AI deployments and interviewed around 350 employees. It highlighted that although popular AI tools like ChatGPT and Copilot have been widely adopted—over 80 percent of companies have explored or piloted them—the gains from these technologies mainly enhance individual productivity rather than impacting profit and loss (P&L) performance.

Only 5 percent of integrated AI pilots generated millions in value, with enterprise-grade AI systems often being quietly rejected by businesses. The report attributes the failure to difficulties in integrating AI with existing workflows and a “learning gap” within the workforce, rather than shortcomings in AI model performance itself.

Executives sometimes mistakenly blame AI’s technical performance, but the real issue lies in adapting organisational processes and training employees to effectively use AI technologies. This insight was echoed recently by Taco Bell’s Chief Digital and Technology Officer, Dane Mathews, who acknowledged slowing down AI rollout in drive-throughs due to inconsistent performance and said that human oversight remains crucial during busy periods.

Mathews emphasized a balanced approach, recommending that teams monitor AI tools carefully and intervene when necessary, underscoring that while AI can boost productivity, human judgment and adaptability remain vital to operational success.