Skills obsolescence: how to know if your team is at risk
The average shelf life of a technical skill has fallen from 30 years in the 1980s to less than 5 years today, according to World Economic Forum data. For digital and AI-related skills, this period can be less than 18 months. An organisation that does not monitor the obsolescence of its skills portfolio takes a measurable operational risk, regardless of its training investment.
What the data say
The World Economic Forum has published its Future of Jobs Report since 2016. The 2025 edition estimates that 44% of workers' skills will be disrupted in the next 5 years, and that the pace of obsolescence is accelerating with the spread of generative AI.
The OECD, in its work on "skills for the future", has documented that the jobs most exposed to automation are not only low-skilled jobs. Some intermediate-skilled white-collar jobs in services (information processing, data entry, standard document analysis) are more exposed than manual jobs requiring dexterity or contextual judgement.
The IBM Institute for Business Value estimated in 2024 that 40% of the global workforce will need to be reskilled in the next 3 years because of AI.
What is the "half-life" of a skill?
The concept of skills half-life comes from research in education science. It denotes the time after which half of what a person learned in a domain becomes obsolete or insufficient to perform at the expected level.
In the 1980s, the half-life of an engineering degree was estimated at 10–15 years. In 2026, it is estimated at less than 5 years for technical skills, and less than 2 years for skills linked to specific digital tools.
This notion has a direct implication for those responsible for training: a training plan that refreshes skills on a 3- to 5-year cycle is structurally behind the obsolescence curve.
How to identify at-risk skills in your organisation
Three criteria assess a skill's exposure to obsolescence.
Degree of routinisation. Skills that consist of applying a rule or procedure to structured data are the most automatable.
Degree of digital mediation. Skills exercised primarily through a screen and keyboard are more easily augmented or replaced than skills requiring physical presence or human relationship.
Regulatory stability. Skills linked to fast-changing regulations or tools age faster than foundational reasoning skills.
What to do
Map skills by half-life, prioritise short half-life skills for continuous learning paths, and maintain human-intensive skills (judgement, relationship, creativity) that are less exposed.