Pillar Guide
30 AI use cases for Swiss SMEs
Concrete, vetted, sorted by ROI × feasibility. No hype theatre.
These use cases come from consulting engagements, industry observation, and workshop results. Not from marketing whitepapers. We sorted them along two axes: ROI (what does it bring?) and feasibility (how hard is it to introduce?). Quick wins at the top, demanding levers at the bottom.
Quick wins (10 use cases)
Low barrier, fast ROI. Typical effort: 5–10 days of Waldsee.
- Mail triage: Classify incoming mail, suggest replies
- Quote preparation: Fill templates automatically
- Meeting minutes: Audio-to-text with summary
- Translation: Internal multilingualism (DE/FR/IT/EN)
- Standard letters: T&C requests, dunning letters, confirmations
- CV screening: First filter in recruiting (not the final decision)
- FAQ bot: Answer internal employee questions from documents
- Excel formula explanation: AI as "the colleague who understands formulas"
- Code review hints: Lightweight, for non-tech SMEs
- Social media captions: Based on images and keywords
Medium levers (10 use cases)
Medium effort, higher ROI. Typical effort: 10–20 days of Waldsee.
- Knowledge bot over internal documents (RAG)
- Automated report generation: Monthly reports
- Lead qualification in the CRM
- Contract analysis: Clause comparison, risk flagging
- Recruiting match: Position ↔ CV pool
- Customer segmentation: With unstructured data
- Revenue/demand forecasting: With an AI layer on existing data
- Helpdesk automation: First-level answers
- Knowledge onboarding: For new employees
- Document classification: In the DMS
Demanding levers (10 use cases)
High effort, transformative. Typical effort: 20+ days of Waldsee plus possibly hardware.
- Code-aware coding copilot on-prem: Tech SMEs
- Anomaly detection in operations data
- Custom GPT for client correspondence: Fiduciary firms
- Own knowledge search: Across Confluence/Notion/SharePoint
- PR review bot: For security-relevant code
- Multi-agent workflows: For complex processing chains
- AI-supported product configuration: In the sales process
- Sentiment analysis: In customer feedback
- Demand forecast: With external market data
- Custom model training: On your own data (rarely worthwhile)
How you turn this list into a strategy
Three steps. First: mark 5 use cases that sound interesting for your operation. Second: book an AI potential assessment. We go through the list, look at what really makes sense for you, and prioritise. Third: you decide. Build it yourself, implement it together, or hand the implementation entirely to us.
Frequently asked questions
Which use case delivers ROI fastest?
Mail triage and quote preparation in most office-heavy SMEs. Code copilots in tech SMEs. Helpdesk automation in service companies.
Do I need my own data for AI use cases?
For quick wins, mostly not. Generic models are enough. For demanding levers (forecasting, custom bots) yes, then data quality matters.
Which use cases are revDSG-critical in Switzerland?
Anything with personal data, client data, patient data, IP-sensitive content. For these use cases we check on-prem or Switzerland-compliant cloud.
What does a typical use case project cost?
Quick wins: 5–10 days of Waldsee effort at the day rate of CHF 2,200. Medium levers: 10–20 days. Demanding levers: 20+ days plus possibly hardware.
Can I build the use cases without Waldsee?
Quick wins yes, with some tech affinity. From medium levers on it gets difficult without experience. Then guidance on a day rate pays off.
Where does the ATHENA framework fit in?
Use case identification belongs in the ATHENA Awareness phase. Implementation in Transformation and Adaptation. More under Methodology.