Companies are pouring billions of dollars into corporate AI projects, but most have yet to see any measurable returns.
A recent MIT report, titled “The GenAI Divide: State of AI in Business 2025,” reveals that while U.S. businesses have collectively invested between $35 billion and $40 billion in AI initiatives, almost all of them (95%) are seeing zero return on their investments or no measurable impact on profits. Only 5% are seeing “value” from AI.
The research, which was based on 150 interviews with AI leaders, an examination of 300 AI applications, and a survey of 350 employees at various companies, found that most AI pilot programs fail to hit targets because of “brittle workflows, lack of contextual learning, and misalignment with day-to-day operations.” In other words, the AI tools do not fit into accepted corporate workflows. Generic tools like ChatGPT “stall” and provide “little to no measurable impact” on profit and loss because they don’t adapt to a company’s established way of doing things, the authors found.
Another key issue is that companies were using AI for the wrong assignments. The research shows that AI works best with back-office tasks with a high return-on-investment (ROI), like administrative and repetitive functions, which many companies outsource. However, more than half of the funds spent on AI projects tried to use the technology for sales and marketing, two areas that the researchers say still need human involvement and have a lower ROI.
The 5% of programs that do succeed in deploying AI seem to focus on one issue. Aditya Challapally, the MIT researcher who led the study, told Fortune that some large companies and younger startups are “excelling” with AI because “they pick one pain point, execute well, and partner smartly with companies who use their tools.” Startups led by young founders have seen revenue “jump from zero to $20 million in a year” following this blueprint, Challapally said.
Additionally, companies that buy AI tools from third-party vendors like OpenAI and Perplexity have an advantage over firms that develop in-house AI tools. The MIT study says that two out of three AI tools from third-party vendors are successful, compared to one-third of in-house tools.
When it comes to AI replacing jobs, the study notes that while there haven’t been AI-related layoffs yet, companies aren’t as quick to replace staff members who leave, especially in customer support and administrative roles. The study states that AI will probably not lead to job loss in the next few years, “until AI systems achieve contextual adaptation and autonomous operation.”
Other AI leaders have had more dire predictions about AI causing job loss. Anthropic CEO Dario Amodei predicted in May that AI could wipe out half of all entry-level, white-collar positions within the next five years.
Join top CEOs, founders and operators at the Level Up conference to unlock strategies for scaling your business, boosting revenue and building sustainable success.
