In this column, guest authors share their views on economic and financial issues.
A new economic paradigm is beginning to emerge—marked by uneven adoption and significant challenges for Europe.
Artificial intelligence inspires as much excitement as it does uncertainty, particularly regarding its potential impact on productivity and employment. Yet AI’s influence on macroeconomic performance ultimately depends on whether companies adopt these technologies on a broad scale. But can we already assess the pace of AI adoption when both the technology and, above all, its practical applications are still in their early stages?
The Federal Reserve Bank of Atlanta (Yotzov et al., 2026) offers one approach by combining several business surveys conducted across four major OECD economies—the United States, Australia, Germany, and the United Kingdom. The objective is to create an internationally comparable, homogeneous, and macroeconomically representative sample. The study is based on a panel of 6,000 companies.
What distinguishes more productive, larger, and higher-paying companies
The survey was conducted between November 2025 and January 2026 and covers companies of all sizes across every major sector of the four economies. The questionnaires were standardized to minimize potential biases. The study identifies four key findings regarding AI adoption.
On average, 69 percent of surveyed companies report using AI, ranging from 78 percent in the United States to 59 percent in Australia. Large language models (LLMs) dominate, accounting for 41 percent of all AI applications, while only 30 percent rely on traditional machine-learning techniques. The findings also show that more productive, larger, and higher-paying companies are more likely to adopt AI than firms led, on average, by older management teams.
Productivity gains will depend on AI investment
At the individual employee level, only 28 percent of respondents report not using AI in the workplace. By contrast, 41 percent of AI users say they use the technology for up to one hour per week, while 24 percent report using it between one and five hours weekly.
In short, many companies have already introduced AI tools. Within organizations, however, adoption remains at an early stage, and the applications currently in use are still relatively basic. The survey also examines expectations for the next three years.
Across the four countries, 75 percent of companies expect to be using AI technologies within the next three years. At the same time, they plan to expand their use of data-driven machine-learning applications. This points to a new wave of investment aimed at building the capabilities and infrastructure required to deploy these technologies effectively.
The study’s findings reinforce the view that we are gradually entering a new economic paradigm. Productivity gains—and ultimately corporate performance—will increasingly depend on companies’ willingness to invest continuously and on the widespread adoption of these technologies across organizations. Europe, however, will first have to overcome a considerable number of structural obstacles.
Mabrouk Chetouane, Head of Global Market Strategy and Romain Aumond, Quantitative Macro Strategist, at Natixis Investment Managers.
- Read all previous guest contributions here.
