The economic impact of generative artificial intelligence is becoming increasingly visible in public economic indicators. In its September 2026 economic monitoring update, the Department of Economic Affairs and the Employment Office of the Canton of Zurich presented concrete figures detailing business adoption and labor market adjustments. Supported by macroeconomic modeling from the Zurich University of Applied Sciences, known as ZHAW, the report presents a distinct dual reality: while generative software provides a substantial boost to value creation, it concurrently amplifies labor market tensions.
According to the survey, 47 percent of surveyed businesses in Zurich now systematically utilize generative AI tools in their daily operations. This represents a substantial rise from 27 percent recorded in 2024. Adoption rates, however, diverge sharply depending on organization size. Large corporations report systematic deployment at 83 percent, compared to 67 percent among mid-sized companies and 45 percent among micro-enterprises.
On an aggregate economic level, the ZHAW baseline model anticipates an additional annual labor productivity growth rate of 1.1 percentage points over the coming decade. Relative to the historical 25-year average of 0.5 percent per year in Zurich, this addition represents more than a doubling of standard gains. The expected productivity surge is heavily concentrated in knowledge-intensive domains. Financial services, insurance, and the information and communication technology sector are projected to realize annual productivity gains of 1.7 to 1.8 percentage points above baseline.
Simultaneously, the monitoring data records pronounced shifts across the employment landscape. Between October 2022 and June 2026, the number of registered job seekers in occupations with high exposure to AI jumped by 118 percent. Over that same timeframe, job seekers in low-exposure roles increased by only 13 percent, while medium-exposure occupations saw an increase of 66 percent. The structural changes primarily impact conventional knowledge roles, including software developers, administrative office personnel, and marketing specialists.
The findings do not point toward blanket workforce contraction, but rather highlight structural polarization. While routine execution in coding and copywriting faces contraction, employment in executive leadership functions and specialized professions, such as architecture, expanded at above-average rates. For cantonal authorities, these figures underline an urgent need for targeted training initiatives to prevent widening labor market mismatches as projected productivity gains materialize.