How to train your teams in AI use at work

Why train your teams on AI now

AI training in the workplace is about giving every employee the skills needed to effectively use generative artificial intelligence tools in their daily work: writing, analysis, research, creation, automation.

In 2026, generative AI is no longer a technological option. It is a productivity tool on the same level as email or spreadsheets. Yet, according to a BCG study (2025), only 28% of European companies have set up a structured AI training programme for their teams.

In French-speaking Switzerland, the picture is similar. Executives invest in licences, but not in training. The result: tools are under-used, employees improvise, and ROI remains unclear.

Step 1: Assess the current level

Before training, you must diagnose. Each employee has a different relationship with AI:

• Enthusiasts (15-20%) already use ChatGPT or Claude daily, sometimes without management knowing.
• The curious (40-50%) have tried it once or twice but do not know how to integrate it into their work.
• The reluctant (30-40%) fear for their jobs or do not see the point.

A 10-minute anonymous questionnaire is enough to map these profiles. This step is essential: identical training for everyone is doomed to fail.

Step 2: Choose the right tools for each profession

The classic mistake: deploying a single tool for the whole company. Each profession has different needs:

• Marketing and communication: assisted writing tools, image generation, GEO optimisation
• Sales: prospecting assistants, account analysis, meeting preparation
• Finance and legal: document summarising, regulatory monitoring, data extraction
• HR: job posting writing, CV analysis, AI-assisted onboarding
• Leadership: AI dashboards, decision support, competitive intelligence

At MCVA Consulting, we recommend starting with 2-3 concrete use cases per department rather than aiming for immediate company-wide adoption.

Step 3: Build a progressive programme

An effective AI training programme runs in three phases:

Phase 1: awareness (weeks 1-2)

Goal: demystify AI and align the team on a shared vision.

• 2-hour workshop: what is generative AI, how does it work, what it can and cannot do
• Live demonstration on concrete cases linked to the company
• Open discussion on fears and expectations

Phase 2: guided practice (weeks 3-6)

Goal: acquire core skills in a safe setting.

• Practical workshops by profession (2 h per week)
• Exercises on real cases from the company
• Advanced prompting techniques: chain of thought, few-shot, format constraints
• Best practices: fact-checking, confidentiality, FADP compliance

Phase 3: autonomy (weeks 7-12)

Goal: integrate AI into daily workflows.

• Each employee identifies 3 tasks they partially delegate to AI
• Sharing sessions: teams present their use cases
• Creation of an internal prompt library

Step 4: Manage resistance to change

Resistance is normal and legitimate. Treating it seriously is essential:

"AI will replace my job." Factual response: AI automates tasks, not whole jobs. An accountant who uses AI to automate data entry has more time for analysis and advice.

"The results are not reliable." True. That is precisely why humans remain necessary. Train people in systematic verification rather than blind trust.

"I don't have time." Training takes time in the short term to gain it in the long term. Studies show productivity gains of 20-40% after 3 months of structured adoption.

Step 5: Measure adoption and ROI

What is not measured is not improved. Set up simple indicators:

• Adoption rate: percentage of employees using AI at least once a week
• Time saved: time saved on recurring tasks (self-reported then measured)
• Quality: impact on the quality of deliverables (internal survey)
• Satisfaction: teams' confidence in using AI

Aim for a 70% adoption rate at 3 months and an average time saving of 5 hours per employee per week at 6 months.

The concrete ROI of AI training

Investing in AI training is not an expense; it is a measurable profitability lever. Available data converges on a clear conclusion.

Documented productivity gains

According to a Harvard Business School / BCG study (2024) of 758 consultants, employees trained in generative AI completed their tasks 25.1% faster and produced results of 40% higher quality than the control group. The gap was even more marked for junior profiles, whose productivity rose by 43%.

A McKinsey Global Institute analysis (2025) estimates that generative AI could automate between 60 and 70% of repetitive tasks in support functions (admin, finance, HR), freeing up time for higher value-added activities.

The cost of not training

Not training also has a price. Companies that deploy AI licences without support typically observe:

• Adoption rate below 20% after 6 months
• Increased compliance risks (confidential data shared with unapproved tools)
• Shadow AI: employees use unsupervised free tools, without governance
• Internal frustration that fuels resistance to change

The average cost of an enterprise AI licence is between CHF 25.– and CHF 50.– per employee per month. Without training, this investment is largely wasted.

Simplified ROI calculation

For a 50-employee company in French-speaking Switzerland:

Even halving the productivity gain to be conservative, the return on investment vastly exceeds the cost of the programme.

An AI competency framework to structure scaling up

Beyond one-off workshops, the companies that succeed in their AI adoption put in place a structured competency framework. Here is a 4-level framework applicable to any profession:

This framework allows you to set clear goals, track progress and recognise the most engaged employees.

Implications for Swiss businesses

The Swiss market presents specifics that make AI training both more necessary and more demanding than elsewhere.

A tight labour market

Switzerland has a structurally low unemployment rate (2.3% in 2025 according to SECO). Recruiting tech profiles remains difficult and costly, especially as the skills that make a difference in the AI era are scarce on the market. Training existing teams on AI use is often more effective, and faster, than seeking to hire specialists.

Specific regulatory requirements

The Federal Act on Data Protection (FADP), in force since September 2023, imposes strict obligations on the processing of personal data. Any AI training in Switzerland must include a compliance component: which data can be submitted to an LLM, which tools are approved, how to anonymise sensitive information.

The European regulatory framework (AI Act) also affects Swiss companies operating in the European market. Anticipating these requirements in training is a competitive advantage.

Multilingualism and cultural proximity

Swiss companies often operate in 2 to 4 languages. Training must cover the specifics of multilingual prompting: when to prompt in English for quality, when to prompt in the target language for cultural nuance, how to manage AI-assisted translations.

SME-driven economy

Switzerland counts more than 600,000 SMEs, representing 99.7% of the country's businesses (FSO, 2024). These structures generally have no internal AI department. An external, pragmatic training programme adapted to their size is often the only realistic way to access these technologies.

Change management tips

Technology represents only 30% of the challenge. The remaining 70% is about the human factor. Here are the practices that make the difference:

1. Involve management from the start. If senior leaders do not use AI themselves, teams will not adopt it either. The signal must come from the top.
2. Celebrate quick wins. As soon as an employee saves time or produces a better result thanks to AI, share it internally. Concrete successes are worth more than speeches.
3. Appoint AI ambassadors. Identify 1 to 2 people per department who become reference points. They bridge formal training and daily use.
4. Create a dedicated channel. A space (Slack, Teams) where employees share prompts, tips and questions. Peer learning is the most powerful driver of adoption.
5. Iterate every quarter. AI tools evolve fast. Plan an update session every 3 months to integrate new features and adjust practices.

The MCVA Executive Training offer

At MCVA Consulting, we offer an Executive AI Training programme specially designed for Swiss companies:

• Initial diagnosis of skills and priority use cases
• Tailor-made programme adapted to your sector and existing tools
• Practical workshops facilitated in French, on site or by video conference
• Post-training follow-up to support lasting adoption

Our courses integrate the specific challenges of the Swiss market: FADP compliance, multilingualism, local sectoral specifics.

Summary

• Only 28% of European companies have a structured AI training programme.
• Start with a diagnosis of levels and needs by profession.
• Roll out a 3-phase programme: awareness, guided practice, autonomy.
• Measure adoption with concrete indicators (usage rate, time saved, quality).
• MCVA Consulting offers Executive AI Training tailored to the Swiss market, contact us to discuss it.

Frequently asked questions

How long does it take to train a team of 20 people on AI?

A complete programme (awareness + guided practice + autonomy) runs over 10 to 12 weeks, at a rate of 2 hours per week. The first results appear from week 4, when employees start applying prompting techniques to their real tasks. For leadership teams, a condensed 2- to 3-day format provides a complete strategic overview.

Is AI training relevant for small businesses (fewer than 20 employees)?

SMEs are often those that benefit most from AI training, as each individual productivity gain has a proportionally bigger impact. A programme tailored to small structures focuses on 3 to 5 priority use cases and can run over 4 to 6 weeks. The cost is quickly amortised: a 3-hour weekly gain for 10 employees represents more than 1,400 hours per year reallocated to value-added tasks.

How do you ensure FADP compliance when using AI in business?

Compliance rests on three pillars: the choice of compliant tools (hosting in Europe or Switzerland, data processing agreements in place), training employees on best practices (never submit non-anonymised personal data), and clear governance (AI use charter, list of approved tools, validation process for new use cases). Our training courses systematically include a module dedicated to FADP compliance.

Want to structure your teams' AI upskilling? Contact MCVA Consulting for a free needs diagnosis, or discover our Executive AI Training programme.