6,000 execs struggle to find the AI productivity boom

Executives report minimal gains from AI

More than 80 percent of corporate executives across the US, UK, Germany, and Australia report no discernible impact from AI on employment or productivity. The National Bureau of Economic Research (NBER) surveyed nearly 6,000 leaders, including CEOs and CFOs, to track how AI is changing the workplace. The findings suggest a massive gap between the hype surrounding generative AI and its actual performance in the corporate world. While tech giants promise a transformation of the global economy, the majority of businesses on the ground have yet to see any measurable change in their bottom line. Most businesses are currently experimenting with the technology even if they are not seeing results. The study shows that 69 percent of companies use some form of AI today, and 75 percent expect to adopt it within the next three years.

Current AI usage focuses on text

Enterprises primarily use large language models for text generation and visual content creation. Data processing using machine learning ranks as the third most common application. These tools have become standard in many offices, yet they have not moved the needle on overall output. The NBER data shows that 90 percent of managers saw no impact on employment levels at their organizations over the last three years. Furthermore, 89 percent of respondents reported no change in productivity, which the study measured as the volume of sales per employee. The lack of progress contradicts the aggressive marketing campaigns from software providers. Many companies are paying for premium AI subscriptions without a clear way to track the return on that investment. The era of experimentation is currently yielding more questions than answers. Common AI applications in the workplace include:

• Text generation using large language models
• Visual content creation for marketing and internal use
• Data processing via traditional machine learning
• Automated coding for software development teams

Job losses loom in future forecasts

While the past three years showed little change, executives expect a shift by 2028. The NBER report estimates that AI will eliminate approximately 1.75 million jobs across the four surveyed nations over the next three years. Respondents also anticipate a modest 1.4 percent increase in productivity due to AI implementation. If this prediction holds true, it would reverse a long-term decline in productivity growth across advanced economies. However, this gain remains small compared to the billions of dollars currently flowing into AI development. Employee expectations do not align with those of senior management. Workers generally expect AI to create more jobs than it destroys over the next three years. They also anticipate smaller productivity gains than their bosses, suggesting a fundamental disagreement on how the technology will affect daily operations.

Corporate leaders struggle with ROI

The NBER paper adds to a growing pile of evidence that AI is not yet living up to its financial promises. A recent PwC survey of 4,500 business leaders found that more than half saw no increase in revenue or decrease in costs from their AI projects. Deloitte reported similar findings in its own research. While 74 percent of organizations want their AI initiatives to drive revenue growth, only 20 percent have actually seen that happen. Most companies are still stuck in the "proof of concept" phase. Specific trials of high-profile tools have also underperformed. A UK government department recently tested Microsoft 365 Copilot and found no overall gain in productivity. The software sped up some tasks but hampered others, resulting in a net-zero impact on efficiency. Key metrics from recent AI surveys include:

• 20 percent: The number of firms seeing revenue growth from AI according to Deloitte
• 50 percent: The portion of leaders seeing zero cost savings in the PwC study
• 91 percent: Customer service leaders feeling pressure to deploy AI
• 84 percent: Leaders planning to reskill agents for AI-integrated roles

Microsoft admits tracking difficulties

Microsoft executives are beginning to acknowledge the difficulty of proving AI's value. Jared Spataro, who leads Microsoft’s AI at Work efforts, recently admitted he struggles to highlight the return on investment for Copilot. He noted that knowledge work does not always translate directly into top-line or bottom-line figures. This admission contrasts with the rhetoric from other Microsoft leaders. Microsoft AI chief Mustafa Suleyman recently claimed that AI will fully automate most tasks involving a computer within the next 12 to 18 months. He specifically cited accounting, legal, marketing, and project management as vulnerable sectors. Other hardware and software vendors continue to push optimistic timelines. Lenovo recently claimed that 94 percent of enterprises in Europe and the Middle East expect a positive return on AI investment soon. These claims often ignore the mounting evidence that real-world integration is slow and difficult.

Customer service faces immediate pressure

The customer service sector is feeling the most intense pressure to automate. A new Gartner survey found that 91 percent of customer service leaders face demands from management to implement AI immediately. Organizations expect AI to handle routine inquiries while humans provide judgment for complex or emotionally sensitive issues. Nearly 80 percent of firms plan to move human agents into new roles as routine tasks become automated. Gartner reports that 84 percent of leaders are already planning to add new skills to agent roles to support this transition. This suggests that while jobs might not disappear immediately, the nature of the work is changing rapidly. The current state of AI adoption reveals a significant disconnect between investment and results:

• Big Tech continues to spend hundreds of billions on infrastructure
• Enterprises are under pressure to deploy tools regardless of proven value
• Productivity gains remain stuck in the low single digits
• Job markets have not yet seen the predicted mass displacement

The technology giants are banking on companies persevering through this period of low returns. For now, the promised massive leap in efficiency remains a future prospect rather than a present reality.