Major IT firms like Microsoft, Google and others are laying off people left and right and replacing them with AI to maximise profit, but a new study suggests that nearly 95 per cent of companies which attempted to incorporate generative AI into their workflow are seeing zero returns.
According to a study conducted by the Massachusetts Institute of Technology titled “The GenAI Divide: State of AI in Business 2025”, U.S.-based firms that have invested somewhere between $35 billion to $40 billion in generative AI are seeing little to no measurable impact on their profit or loss.
The research, which included 150 interviews with AI leaders, an examination of 300 AI applications, and a survey of 350 employees at different organisations, found that the majority of AI pilot programs failed to meet their targets because of “brittle workflows, lack of contextual learning, and misalignment with day-to-day operations.”
And while AI did work for back office tasks like administrative and repetitive functions, more than half of the money spent on AI was spent on sales and marketing and failed, two areas where researchers say human involvement is still required.
In a statement to Fortune, Aditya Challlapally, the lead author of the report, said that while some companies “have seen revenues jump from zero to $20 million in a year”, the majority of organisations are having a hard time as there is a huge “learning gap.”
Despite executives blaming regulations and model performance, the study revealed that the problem lay in the flawed enterprise integration of AI tools. Researchers also noted that consumer-centric AI chatbots like ChatGPT often stalled and had little to no impact on the output.
What the 5 per cent of firms using AI for real growth did right
The major difference between companies that successfully used AI to drive growth and those that didn’t lies in the adoption and implementation of the technology. Firms that purchased AI tools from specialised vendors succeeded around 67 per cent of the time, while companies that developed in-house products saw a success of just 33 per cent.
Story continues below this ad
MIT’s new study also shows that companies that went solo had higher failure rates than those that purchased customised solutions. Also, it was noted that managers, driving adoption and using tools that adapt over time, helped drive growth and productivity.
The study also highlights how companies are trying out agentic AI systems to learn, remember and perform certain actions, giving us a glimpse of how enterprise AI will evolve over time.
Expand
© IE Online Media Services Pvt Ltd