MIT study: 95% GenAI projects fail to show returns

While companies have rushed to adopt AI tools into their work processes, most of these projects haven’t made any impact as only 5% of AI pilots showed growth in revenue. 

A new MIT study, titled ‘The GenAI Divide: State of AI in Business 2025,’ which surveyed 300 AI deployments and spoke to 350 employees, found that 95% of AI projects showed no returns on investment for the companies. 

The failure was not because AI models did not work efficiently, but because they were harder to adapt with pre-existing workflows in a company. 

According to a Fortune report, citing the study’s lead researcher, the 5% AI deployments that were successful were focused on solving a single problem and had worked on execution. 

Companies are also dealing with a “learning gap” in their workforce although company executives blamed AI model’s performance. 

The study also showed that more than half of GenAI budgets went to sales and marketing but the highest returns were seen in automating back-office processes like cutting external agency costs, and streamlining roles.

The report also noted that companies that purchasing AI tools from specialised vendors worked around 67% of the time while internal tools only succeeded one-third of the time.