AI: Waiting is not a strategy (and even less an excuse)
Remember COVID
The famous maxim"As fast as possible, but as slow as necessary" made perfect sense at the time.
But when applied to AI in 2026, it becomes a real strategic oxymoron.
Because while you hesitate, your employees are already using AI, sometimes without you even knowing it.
Welcome to shadow AI —the use of AI tools via personal accounts, with the volume of data sent to these services increasing sixfold in one year.
And that's not all: nearly half of business users access AI through credentials that are not managed by the organization.
In other words: the risk is not in AI... it's in your lack of an AI strategy.
Nearly half a century at the heart of the economic fabric of SMEs in French-speaking Switzerland has taught us one essential truth:technology only has value when it is used to serve people. It is this deep conviction that guides every action we take, every project we undertake, and every solution we deploy.
Three mistakes we see everywhere
1. "We'll put that in IT."
Entrusting "AI" to IT is like asking your accountant to decide on an acquisition.
AI is primarily a matter of strategy, business, governance, risk, and human capital.
Swiss regulators themselves (FINMA, for example) point out that AI requires management that includes:
- governance,
- inventory of use cases,
- explainability,
- human control,
- documentation,
- model risk management.
IT executes.
Management decides.
2. "The solutions are not mature enough / our data is not clean enough."
Swiss SMEs are already making progress: by 2025, 34% will have integrated AI into their processes (vs. 22% in 2024).
The paradox?
Only one in three SMEs has implemented internal data protection rules (and only 23% of the smallest ones).
In other words: technological maturity is not the problem.
The problem is the framework.
And in Switzerland, the framework exists:
- The nLPD is directly applicable to AI: transparency, individual rights, impact assessments, human oversight of automated decisions.
- The Confederation has published AI guidelines: consistency, responsibility, robustness, and future sectoral reinforcement.
- The EU is rolling out the AI Act (mandatory for companies operating in the EU).
In short: what is lacking is not maturity, but decision.
3. "Yes, but we have other urgent matters to attend to."
The data is clear: companies that are truly successful with AI are not those that "experiment" with it, but those that industrialize AI in their workflows.
However, according to McKinsey, 88% of companies already use AI, but only a third manage to deploy it effectively, and 6% derive real business performance from it.
The obstacle? The inability to move from perpetual POC to actual deployment.
Waiting gives the most agile players a lasting competitive advantage.
The real risks... are those of inaction.
1. Losing your competitive advantage
The adoption of AI is progressing rapidly in Switzerland, both in SMEs and in the public sector.
Failure to keep up means exposing oneself to an almost unbridgeable gap in productivity, service quality, and speed of execution.
2. Let the shadow AI explode
When nearly half of internal AI usage goes through personal accounts, you potentially expose yourself to:
- sensitive data,
- customer information,
- trade secrets,
- internal identifiers.
[cio-online.com]
3. Reduce the employability of your employees
OECD analyses show that productivity potential can only be realized if organizations develop their teams' AI skills.
Without this, there will be a loss of employability, efficiency, and regional attractiveness.
The figures (Switzerland & French-speaking Switzerland)
- 34% of Swiss SMEs will be using AI in 2025 (vs. 22% in 2024).
- Only 33% have defined internal rules on data.
- In French-speaking Switzerland, 895 professionals were surveyed:
- 55% use it for administrative purposes,
- 33% for marketing,
- 25% for R&D,
- 32% do not yet use AI.
- In 2025, 60% of the Swiss population uses AI tools (vs. 40% the previous year).
Mass adoption among those under 35: 79%.
Productivity: promises, realities, and methodology
The OECD 2024 study shows that AI can boost productivity if it is integrated into processes, not just grafted onto existing tools.
Successful organizations:
- select 3–5 use cases with a high P&L impact,
- rethink processes around AI,
- measure the effects,
- train intensively.
It's not a question of "technology."
It's a question of organizational design.
The responsibility of the manager
The Swiss authorities are very clear:
- AI governance,
- transparency,
- risk management,
- human control,
- documentation,
- ongoing supervision.
AI is not a subject that can be delegated to IT to "see what they can do."
It is a matter of leadership, risk, competitiveness, public service, and collective responsibility.
Action plan (90 days)
0 — Stop shadow AI (Weeks 1–2)
- Immediate usage policy: authorized/prohibited data, approved tools, transparency requirements.
- Mandatory business accounts, discontinuation of personal accounts.
- “Safe & Effective AI” training (half day).
1 — Define the ambition and use cases (Weeks 1–4)
- COMEX workshop: 3 measurable objectives (e.g., reduction in cycle time, commercial gains, improvement in support).
- Use case registry: value, risks, data, complexity, ROI.
2 — Establish governance (Weeks 2–6)
- Comprehensive AI policy.
- Risk register / impact analysis of the Data Protection Act.
- Pre-AI Act alignment for companies operating in the EU.
3 — Deploy & measure (Weeks 4–12)
- Re-design process.
- KPIs (speed, cost, quality, conversion, actual productivity).
- Continuous review (bias, errors, robustness).
4 — Provide extensive training (ongoing)
OECD studies show that AI gains are proportional to skill development.
A word to public officials
The Swiss Confederation has a digital administration strategy for 2024–2027 and AI guidelines.
The role of municipalities, cantons, and administrations:
- modernize processes,
- improve service quality,
- optimize the use of public funds,
- equip the agents.
Conclusion: the real decision to be made
Until the day when:
- your talents surpass you,
- your competitors are pulling ahead of you,
- your customers are leaving you,
- Your public services are falling behind.