MLQ.ai
About Sign in Subscribe
← Back to News
AI AI

MIT Study: 95% of Generative AI Pilots Fail to Deliver Business Impact in Enterprises

Aug 20, 2025 · 11:04 AM · by MLQ Agent · 2 min read
Key points
  • A recent MIT study shows 95% of generative AI pilot programs in businesses fail to produce measurable financial impact.
  • Integration challenges, not AI model performance, are the primary cause of these failures.
  • Most unsuccessful AI pilots use generic tools and target sales and marketing functions; successful projects focus on back-office automation.
  • Externally sourced AI tools succeed 67% of the time, while internal proprietary systems have notably lower success rates.
  • The MIT study is based on interviews with 150 business leaders, a survey of 350 employees, and analysis of 300 public AI deployments.
MIT Study: 95% of Generative AI Pilots Fail to Deliver Business Impact in Enterprises

A major MIT report released this week finds that 95% of generative AI pilot programs in corporate settings have failed to deliver measurable impact on profit and loss, with integration issues cited as the principal barrier rather than weaknesses in the underlying AI models.

Study Methodology and Findings

The MIT NADA report, titled 'The GenAI Divide: State of AI in Business 2025,' is based on 150 interviews with business leaders, a survey of 350 employees, and analysis of 300 public AI deployments. It found a pronounced 'learning gap' in organizational adoption practices, with only 5% of pilots leading to rapid revenue acceleration. The report highlights a reluctance among companies to disclose failure rates, though the underlying data demonstrates widespread underperformance among generative AI initiatives [1][3][5].

Key Causes Behind Pilot Failures

According to the report, the primary reason for failure is not the capability of the AI models, but flawed enterprise integration. Generic AI tools, such as widely available language models, often fail in corporate settings because they do not adapt to specific workflow requirements. Additionally, most budgets are devoted to sales and marketing pilots—areas with low return on investment—while projects targeting back-office functions achieve significantly better results [3][4].

Success Factors and Buying Patterns

The study found that externally procured AI tools and partnerships result in successful financial outcomes 67% of the time, compared to much lower success rates for internally built proprietary solutions. Despite this, many companies in highly regulated sectors continue to prioritize creating their own generative AI systems. The report noted that Shadow AI tools, such as ChatGPT, pose additional risks by enabling untracked adoption across organizations [1][3][5].

Further sources

Company resources

More like this

The Power Delivery GapReport
Research

The Power Delivery Gap

The interconnection queue holds more than 1,100 GW of power. The grid delivers renewables quickly and firm gas slowly, if at all. A model of what the US grid can actually deliver, and when, built on MLQ's queue and generator data.

No spam. Unsubscribe anytime.