Observatory · Switzerland · 2025 Edition
Data & Artificial
Intelligence
Measuring the acceleration — from experimentation to responsible industrialisation. A synthesis of the key AI transformation trends among Swiss organisations.
Key transformation indicators
At-scale deployments
52%
↑ +31 pts vs 2024
Leaders with “good to very good” AI knowledge
62%
↑ +25 pts vs 2024
Low or intermediate maturity
70%
↓ −21 pts vs 2024
AI integrated into strategic plan
35%
↑ +2 pts vs 2024
Data quality “good to excellent”
62%
↑ +14 pts vs 2024
ROI measurement “basic or non-existent”
58%
↓ −16 pts vs 2024
AI understanding among executives
Phase of AI initiatives
Three structural shifts
What has truly changed in a year
Between 2024 and 2025, three structural dynamics are reshaping the AI landscape in Switzerland.
Accelerated industrialisation
At-scale deployments reach 52%, while proof-of-concepts fall to 39% (−28 pts). AI strategies are finally leaving the lab to take root in operational realities.
Expanded governance
An AI project is now assessed as much on its profitability as on its ethical or energy footprint. 58% still rate ROI measurement as basic or non-existent — down from 74% in 2024, a notable improvement.
Human capability in transition
The share of executives with “good to very good” AI knowledge jumps from 37% to 62%. Yet 7 in 10 organisations still identify culture and organisation as the primary barrier.
“A data-driven culture boosts AI expertise: each step forward in data maturity nearly doubles the likelihood of reaching a high level of AI expertise.”
Key barriers identified
What is still holding back progress
Despite the acceleration, four structural obstacles persist and shape the success of the transformation.
Culture and organisation (70%)
Among non-AI users, organisational obstacles are a greater barrier (65%) than technical limitations (30%). Change management must be the top priority.
Value measurement still insufficient
58% of organisations have a “basic or non-existent” AI ROI measurement. Projects track activity metrics (number of models deployed) rather than real business impact.
Sustainability & sovereignty under-addressed
72% acknowledge that sustainability requirements are not properly addressed — up 5 pts vs 2024. “Green AI” is gaining traction but still lacks a standardised methodological framework.
Strategic knowing-doing gap
74% believe AI can solve their main challenges, but only 35% integrate it into their strategic plan — a 39-point gap revealing a persistent organisational wall.
Generative AI spotlight
From excitement to operational adoption
In one year, generative AI has crossed a decisive threshold, moving from pilot to large-scale production.
Deployed “at scale”
52%
↑ from 21% in 2024
See a deep transformation of their role
35%
↑ from 19% in 2024
Customer / marketing / after-sales use cases
42%
↓ from 51% — spreading to back-office functions
Use cases by domain
AI in every function
A selection of the key use cases identified in 2024 and 2025, by discipline.
| Domain | Representative use cases |
|---|---|
| Finance | Data volume synthesis, anomaly and fraud detection, supplier categorisation, dynamic discounts, spend classification. |
| HR | Skills frameworks, personalised training paths, recruitment automation. New 2025: job offer writing and distribution via generative AI. |
| Operations | Quality control via image analysis, failure and delay prediction, resource planning, supply/demand forecasting. |
| CIO / IT | Code generation assistance, IT request classification, process documentation, helpdesk workload management. |
| Customer relations | Automated digital assistants, 360° customer view, Next Best Action/Offer, meeting summaries, opportunity prediction. New 2025: AI agents on customer service (e.g. Wiley). |
| Legal | Case law research, contract synthesis, litigation success estimation, global regulation monitoring. |
| Logistics | New 2025: real-time defect detection on parts (Innodura), transport optimisation via AI models (CMA CGM). |
| Media | New 2025: copyright protection and archive monetisation (Le Monde, AFP × Mistral AI; Time, Fortune × Perplexity AI). |
Strategic recommendations
Three decisive levers for 2025–2026
Organisations that will scale sustainably share a common blueprint built around these three pillars.
01
Explicit strategic vision
Link every AI use case to growth, profitability and ESG objectives. A single executive sponsor (CEO/CFO) + P&L mandate for the CDAO — not mere compliance.
02
Consolidated organisation
Interoperable platforms, robust MLOps practices, governance shared across business, IT and compliance. Monthly value committees crossing operational KPIs with business metrics.
03
Systematic AI acculturation
“AI Champions” embedded in every department. Over 60% of budgets dedicated to prompt training, process redesign and internal communication.
“The race is no longer to the proof of concept; it is to responsible industrialisation. Those who succeed will bring together strategic vision, operational excellence and societal awareness to turn AI potential into lasting competitive advantage.”
