95% of Businesses Report Zero Returns on In-House AI, MIT Study Shows

Investments September 6, 2025

U.S. companies have funneled an estimated $35 to $40 billion into internal AI projects. Yet according to a new report from MIT, nearly 95% of those businesses have failed to generate profit or measurable impact from those efforts. Only 5% report any real value.

The findings came from The GenAI Divide: State of AI in Business 2025, a study that looked at 300 AI apps, conducted 150 interviews with industry experts, and surveyed 350 employees.

The report paints a clear picture: most AI pilots fail because they don’t fit into existing workflows. Researchers found that generic tools, such as widely used chatbots, often stall out because they struggle to adapt to company-specific operations.

Why AI Misses the Mark

Freepik | Businesses often expect AI to succeed in areas where it adds little real value.

One of the biggest issues lies in how businesses apply AI. The MIT study noted that while AI works well for back-office functions like administrative and repetitive tasks, companies keep pouring money into sales and marketing automation, which often deliver weaker returns and still depend heavily on human input.

The report also highlighted common barriers, including:

1. Rigid workflows that don’t allow AI tools to adapt
2. Limited contextual learning, leaving tools unable to adjust to unique challenges
3. Mismatch with daily operations, preventing meaningful impact

In short, companies often expect AI to transform areas where it adds little value.

The 5% That Succeed

While most firms miss out, a small group thrives. These successful adopters tend to focus their AI projects on solving one specific problem.

MIT researcher Aditya Challapally, who led the study, explained that standout companies and startups often choose a single pain point and execute well, sometimes partnering with external providers who specialize in AI tools. He noted that in certain cases, young startups boosted revenue “from zero to $20 million within a year” by following this targeted approach.

This success rate contrasts with the struggles of companies developing their own systems. According to the report, two out of three third-party AI solutions produce favorable results, compared to one out of every three in-house efforts.

Will AI Replace Jobs?

Freepik | Widespread job disruption is unlikely as long as AI lacks the ability to make its own complex decisions.

The MIT report also addressed growing concerns about automation and layoffs. While researchers found no widespread job loss so far, they did observe that some companies have slowed down hiring replacements, especially in customer support and administrative roles.

The study concluded that significant job disruption is unlikely until AI systems achieve advanced contextual adaptation and autonomous decision-making. However, industry leaders hold different views. Earlier this year, Anthropic CEO Dario Amodei predicted that AI could potentially replace half of all entry-level white-collar jobs within five years.

What Businesses Can Learn

The takeaway is clear: AI isn’t a guaranteed shortcut to growth. Companies that expect sweeping change often walk away disappointed. The ones that succeed take a narrower, more strategic approach—focusing on high-impact tasks, integrating tools that fit seamlessly into operations, and relying on proven third-party solutions instead of building everything internally.

With AI investment showing no signs of slowing down, the difference between wasted spending and measurable success will come down to alignment, focus, and execution.