90% of Companies Say AI Is Paying Off — But Only 2% Have Actually Cut Jobs Because of It

Corporate boardroom with AI ROI dashboard

The AI ROI Reality Check

US companies spent $37 billion on generative AI alone in 2025. Boards and executives are now asking the obvious question: is any of it paying off? A major new survey from Harvard Business Review provides some surprisingly clear answers.

The study, based on interviews with 12 enterprise AI leaders and a survey of 1,006 global senior executives, found that 90% of respondents say AI is delivering value — 45% report "a great deal" of value, and another 45% report "moderate" value. Only 9% said their organizations are seeing small returns.

The Inconvenient Truth About AI and Jobs

Here is the number that should stop the "AI is replacing everyone" panic: only 2% of announced headcount reductions are actually enabled by existing AI capabilities in production. The rest is either anticipatory — companies cutting jobs hoping AI will fill the gap eventually — or outright "AI washing" of reductions made for entirely different reasons.

In other words, when companies announce layoffs and blame AI, they are mostly lying. The technology is not yet capable of replacing most workers in practice, even if it might theoretically be able to in the future.

The 7 Factors That Actually Drive AI Value

The real insight from the HBR study is not whether AI pays off — it clearly does for most companies. The insight is how the high-value companies differ from the rest. And the answer is not about technology sophistication:

  1. Clear value definition — leaders explicitly define what "value" means for their organization, whether that is short-term ROI or long-term transformation
  2. Accountability — specific people are held responsible for AI outcomes
  3. Measurement discipline — treating AI measurement as a management practice, not a reporting exercise
  4. Product-focused AI — looking beyond internal process improvements to AI in customer-facing products
  5. Modern tech infrastructure — companies built on modern tech stacks unlock more value
  6. Proprietary data investment — organizations deeply invested in their own data see better returns
  7. AI at the center — putting AI at the core of business transformation, not treating it as an add-on

The Budget Freeze Threat

There is real urgency behind these findings: 71% of global CIOs say their AI budgets will be frozen or cut if value cannot be demonstrated within two years. The honeymoon period for AI spending is ending, and executives who cannot show results are going to lose their funding.

This explains why the high-value companies are so focused on measurement and accountability. When the CFO asks "what are we getting for our $37 billion?" the answer needs to be specific, quantifiable, and tied to business outcomes.

The Bottom Line

The HBR study paints a picture that is both more optimistic and more nuanced than the typical AI narrative. Yes, AI is paying off for most companies. No, it is not replacing workers at scale. And the companies getting the most value are not the ones with the most advanced technology — they are the ones with the most disciplined approach to defining, measuring, and being accountable for AI outcomes.

For anyone worried about AI taking their job: relax, at least for now. For anyone spending millions on AI without clear metrics: start worrying. The budget freeze is coming.