AI Skills at Work: What China Teaches While Swiss SMEs Still Debate

The Global Benchmark: China Teaches AI Starting in Primary School

In May 2025, China's Ministry of Education approved a tiered AI curriculum that became mandatory nationwide from September 2025. Primary school pupils now receive at least six hours of foundational AI awareness per year, middle school students add ten hours of AI logic, and high school students spend one hour per week on AI innovation. The stated goal: by 2027, every student in the country should be AI-literate. In Beijing alone, more than 1,400 schools started offering at least eight hours of AI education from autumn 2025, fully integrated into existing IT curricula.

Why This Is More Than a Footnote for Swiss SMEs

One could argue that national curricula have little to do with the daily business of a Swiss mid-market company. But the real point lies elsewhere: an entire generation of future employees, customers and competitors is growing up treating AI as a given, not a specialty topic. Companies still debating basic policy questions today aren't losing ground to a better language model — they're losing ground to organisations that have already learned to work with AI. For a closer look at how this global race plays out and which mistakes to avoid, see avoiding the three biggest mistakes in the global AI race.

The Swiss Reality: High Priority, Little Plan

Digitalswitzerland surveyed DACH SMEs on the state of upskilling in 2025. The result is revealing: 78% of companies rank AI training for employees as a priority. Yet only 23% have a concrete plan for how to get there. The gap between intention and execution is larger than the gap between willingness and knowledge — and closing it starts with a structured approach rather than another round of internal debate.

What Structured AI Training Actually Delivers

Numbers from markets further along the curve tell a clear story. In the UK, the government's Skills for AI initiative found that 50% of companies report measurably higher productivity from targeted AI training, while 26% see new responsibilities emerge within their teams as a result. The OECD's D4SME report notes that 21% of SMEs already use generative AI specifically to compensate for skill gaps in their workforce — evidence that tools and training belong together.

The gap between tool adoption and actual competence becomes especially clear in a survey by NAB: 58% of SMEs already see higher productivity from AI, yet only 17% of their workforce feel genuinely ready for it — rating their own readiness at an average of 3.8 out of 10. Investment in tools alone doesn't explain success; what matters is whether employees can actually work with them.

AI is a management topic, not an IT project. Without systematic training for employees, the return on investment simply doesn't materialise — no matter how good the underlying model is.

Action Areas for Systematic Upskilling

The experience of leading companies points to a consistent pattern that translates well to Swiss SMEs. What matters is not any single measure, but the interplay of several action areas:

1. Structured training instead of informal experimentation — a clear curriculum rather than solo learning-by-doing.
2. Hands-on and practice-oriented rather than pure awareness sessions — employees work on real cases from their own daily tasks.
3. Role-specific formats for sales, support, HR and finance — application differs significantly by function.
4. Embedding in existing processes rather than one-off training — AI use becomes part of daily work, not an add-on task.
5. Change management alongside training — new tools shift roles, workflows and sometimes status structures within teams.
6. Building internal AI champions — early adopters who act as multipliers and bring colleagues along in everyday work.

One action area is often underestimated: compliance and governance belong in training from day one, not as an afterthought. The EU AI Act and Switzerland's revised Data Protection Act define guardrails that employees need to know before they use AI tools with customer contact or personal data. Getting this foundation right early prevents costly rework later.

That binding guidelines and training initiatives also work in traditional industries is demonstrated by Funke Mediengruppe, which has systematically embedded AI competence into its newsrooms and educational programmes. For a closer look at how this transformation unfolds in practice, see how AI is transforming journalism and education.

The First Step Matters More Than the Perfect Plan

No Swiss SME needs to replicate China's education system or build its own AI academy to stay competitive. What matters is a realistic, role-specific starting point that delivers visible results within a few weeks — and can then be scaled systematically. This is exactly where building an external AI capability centre comes in: one that trains your team, guides implementation and puts the right guardrails in place, without requiring you to build an in-house team from scratch.