Even as AI investment is surging, the bigger question is whether enterprises are becoming more productive—or simply doing the same work faster.
AI is beating all records for the sheer amount of investment it has attracted, and a recent Reuters analysis highlighted the AI investment boom. Global investment in AI infrastructure could reach $31.6 trillion by 2050, with data centres accounting for much of that spending, according to PWC. Meanwhile, AI companies themselves are committing extraordinary sums to computing infrastructure.
Underneath the scale of investment, economists are asking the obvious question: where is the productivity payoff? The answer, emerging from a growing body of research, is more complicated than the AI boom might suggest.
There is little doubt that AI can make individual workers more productive. A 2023 National Bureau of Economic Research (NBER) study of more than 5,000 customer-support agents found that access to a generative AI assistant increased productivity by nearly 14%, with gains reaching 34% among less experienced and lower-skilled workers.
But making an individual task faster is not the same as transforming an organisation. That distinction is becoming increasingly important. McKinsey’s 2026 analysis of AI productivity describes an emerging paradox: AI adoption and investment are accelerating, yet sustained business impact remains elusive. Almost nine in ten companies had deployed AI in at least one business function by the end of 2025, but 94% of respondents said they had not seen significant value from those investments.
The McKinsey study finds that many companies remain focused on incremental productivity gains rather than reshaping products, business models and operating structures.
Research into how people actually use AI at work reinforces the point. An NBER study on the impact of Gen AI in 2025 comprising 7,137 knowledge workers in 66 firms found that while AI reduced time spent on email and helped workers complete documents faster. But it did not significantly change time spent in meetings or the overall quantity and composition of tasks. In other words, workers became more efficient at parts of their existing jobs, while the broader organisation largely stayed the same.
That may be the real AI productivity paradox: that AI can accelerate the existing enterprise without necessarily reinventing it. Microsoft’s 2026 Work Trend Index suggests that AI’s impact depends as much on the organisation as on the individual. Drawing on data from 20,000 AI users across 10 countries, the study found that culture, management support and talent practices had twice the reported impact than individual effort alone.
Only 13% of AI users said they were rewarded for reinventing work with AI, while 45% said it felt safer to focus on existing goals than redesign work around AI. Microsoft calls this the “Transformation Paradox.”
The implication for CIOs and business leaders is significant. The question should no longer be how much time AI saved, as much as much as what can the organisation do differently because the time that was saved.
The brings into focus more crucial questions to examine such as, did the introduction of AI eliminate unnecessary processes or helped redesign roles by removing layers of approval. Did AI helped create new products, helped changed customer experiences. Then examine a more likely scenario whether employees simply used the saved time to handle more emails, attend the same meetings to produce more of the same output. Industry watchers are asking tough questions about the productivity output of AI investments and putting pressure that the next phase of enterprise AI should be less about adoption and more about redesign. This will create a new class of winning companies that are willing to rethink the workflows and not necessarily those that put AI into the most workflows.
About The Author
