ZURICH - The Deloitte analysis crosses two datasets: the demographic replacement demand in 279 Swiss occupations, and a newly developed index of how far today's generative AI systems can transform each job. The result is a mismatch at the heart of the labour market. The occupations ageing fastest are precisely those where AI helps least: nursing and care, teaching, the skilled trades and transport.
The figures behind that conclusion are stark. Among qualified nursing professionals, just under 22,000 people, 23 percent of the occupation, are 55 or older. In healthcare support roles the share reaches 30 percent, among bus and tram drivers 35 percent, and among domestic helpers and cleaners 33 percent. Education and teaching occupations also show above-average replacement demand. AI can make documentation, planning, administration and information retrieval easier in all of these jobs. It cannot, for the foreseeable future, care for the elderly, teach children or carry out physical labour.
The risk is not replacement, it is laziness
Antonio Russo, innovation leader at Deloitte Switzerland, turns the usual fear on its head. The main danger is not that AI will replace jobs, he argues; it is that companies will leave their jobs unchanged and simply overlay them with AI. Anyone who uses the technology merely to write emails faster or automate administrative tasks is wasting most of its potential. The real opportunity lies in redesigning entire business processes and asking which tasks will still need a human at all.
Michael Grampp, Deloitte's chief economist in Switzerland, puts the policy version of the same warning: the country must not mistake AI for a simple solution to the demographic labour shortage. Generative AI transforms knowledge-intensive and administrative tasks, he says, while the high replacement demand sits in occupations built on physical presence, manual skill, human judgement and direct interaction.
Where AI does bite: entry-level office work
The study finds one quadrant where the technology's potential and the demographic pressure do align badly for workers: software and marketing occupations couple high AI transformation potential with low replacement demand. Job cuts cannot be ruled out there, the authors write, and entry-level roles could disappear, the very positions through which young people usually enter these industries.
For everyone else, the prescription is old-fashioned: training, career paths and better working conditions in the shortage occupations, plus a serious attempt to redesign work around the technology rather than sprinkle it on top. The retirement wave, the study notes, is coming either way. The labour market data already points the same direction, with trades and technicians topping the autumn demand lists and canton Zurich's unemployment steady at 2.9 percent, as parliament weighs more flexible working time rules to make the remaining workforce stretch further.