Methodology
Methodology for AI adoption in Swiss SMEs
Classically structured: analysis, prioritisation, pilot, enablement, operations and scaling. Swiss, no buzzword theatre.
The classical implementation logic
The methodology is deliberately classical. First understand, then prioritise, then pilot, enable and move into operations. Each step answers a question a Swiss SME realistically asks itself. Plus a Waldsee service that steps in exactly there.
| Step | What | Typical question | Service |
|---|---|---|---|
| 1 · Analysis | Understand opportunities, risks and the current state | Where does AI even make sense for us? | AI Potential Assessment |
| 2 · Prioritisation | Evaluate use cases by value, effort, data and risk | How do we trust the tech without being naive? | Strategy Workshop |
| 3 · Process design | Bring people, process and tool together cleanly | How do we keep employees at the centre? | Implementation Guidance |
| 4 · Pilot & enablement | Build the solution and enable the team | How does the tool actually get used? | Waldsee Training |
| 5 · Operations | Hold the course, adjust and secure governance | How do we keep learning without it becoming a project? | Day-Rate Implementation |
| 6 · Scaling | Standardise and expand measurable results | What do we do with what works? | Implementation |
Context engineering
The hype term in 2024 was "prompt engineering". By now the industry knows: a good prompt alone is not enough. What really counts is the context an AI tool works in. The data, the connections, the permissions, the knowledge base, the workflows around it.
We call this context engineering: not "find the right trick" but "build the right system". This is why the same ChatGPT licence performs 100× differently in two companies. The model is not different. The context is different.
Use-case evaluation
An AI use case is not automatically good just because it is technically possible. We evaluate use cases by classical criteria before they move into implementation:
| Criterion | Guiding question |
|---|---|
| Value | What economic or operational leverage does it create? |
| Feasibility | Is the use case realistic with the team, budget and system landscape? |
| Data readiness | Are data, documents and permissions clean enough? |
| Risk | Which privacy, security or governance questions arise? |
| Operations | Who will use, maintain and improve the solution after the pilot? |
We prioritise only what brings value, feasibility and operations together. That creates a system for daily work, not a demo.
How the building blocks work together
The methodology provides the sequence. Context engineering provides the technical working mode: systems instead of prompt tricks. Use-case evaluation provides the selection criteria. Together they form a pragmatic approach that starts quickly and is still robust enough for operations.
Common questions
What role does ATHENA still play?
ATHENA remains an internal shorthand and a structure in the executive curriculum. On the website, we deliberately present the methodology in classical terms: analysis, prioritisation, pilot, enablement, operations and scaling.
Can I apply the methodology myself?
Yes. That is the point: we do not sell a consulting secret. Most Swiss SMEs can handle the first steps themselves and bring in support where prioritisation, architecture or implementation become demanding.
What if we already use OKR or PRINCE2?
Then the methodology docks onto it. It does not replace an existing management or project logic. It translates AI adoption into clear decision points.
Why classical methodology in an AI consultancy?
Because AI projects rarely fail from lack of hype. They fail because goals are unclear, data is messy, acceptance is missing and nobody owns operations.
The methodology in practice
Start with the Awareness phase: a structured finding of where AI makes economic sense in your operation.