Implement · Phases T/E/A
AI implementation for Swiss companies
We build what works: apps, dashboards, automations, on-prem setups. Hands-on, at the day rate, without lock-in.
- Hands-on, no slide theatre
- CHF 2,200 / day
- On-prem capable (ThinkStation PGX)
- n8n, Custom GPT, Ollama, Power BI
- ATHENA phases T/E/A
Four delivery models
App / Workflow
App implementation
Custom GPTs, ChatGPT Enterprise, Copilot, n8n workflows.
Data
Dashboard building
Reporting, monitoring, forecasting with an AI layer.
Compliance
On-premise AI
Lenovo ThinkStation PGX at your premises, revDSG-aware.
Guidance
Implementation guidance
Your team builds, we are at your side. Day rate or retainer.
You know where AI helps. Maybe you had an audit. Maybe your team researched it themselves. Now you need someone who puts the thing in place.
No PowerPoint with a phase diagram. Working software in your business. We build hands-on: custom GPTs for your knowledge base, n8n workflows for mail triage and document flow, Power BI or Looker dashboards with an AI layer, on-prem LLM setups when data must not leave the business. Day rate CHF 2,200. You see what you pay. And you see what comes out.
ATHENA context
In the ATHENA methodology, implementation falls into the phases Transformation (from process to workflow), Empowerment (enabling the team to operate the tool) and Adaptation (adjusting the setup to real usage). We work in that order. Not backwards.
Use-case examples
- Tech SME: coding copilot local on the code repository → For tech teams
- Classic SME: mail triage with n8n + Custom GPT → For SMEs
- CXO: top-down AI strategy + implementation guidance → For CXO
Frequently asked questions
How long does an AI implementation take?
It depends on the use case. An n8n mail triage is up and running in 2–5 days. A custom GPT with RAG connection in 1–3 weeks. An on-prem setup on a ThinkStation PGX in 2–6 weeks, depending on hardware lead time.
Does Waldsee implement itself, or only advise?
Both, you choose. We build directly (day rate CHF 2,200), or your team builds and we are at your side. The guided variant is often cheaper and more sustainable if you have technical people in-house.
Which tools do you use by default?
n8n for workflow automation, Custom GPT / Claude Projects for knowledge apps, Ollama + Open-WebUI for on-prem, Power BI / Looker Studio for dashboards. The stack leans open-source and is swappable.
What does an implementation cost in total?
Day rate × days + (for on-prem) hardware. A typical workflow-automation project: 5–15 days of Waldsee effort. An on-prem setup: 3–8 days of Waldsee plus ThinkStation hardware.
Do I need an audit first?
Not necessarily. If you already know the use cases, we start directly. If you are unsure where the leverage is: audit first (CHF 5,000–15,000 fixed price), then implementation.
What happens if a tool no longer fits later?
An open-source stack means: swappable. A model, a workflow, a UI: all can be replaced without lock-in. Documentation is part of the delivery standard.
Can I book guidance only, without a build mandate?
Yes. Guidance at the day rate, possible in half-days. Often the most efficient variant if you already have a tech lead in-house.
Which industries do you support?
Focus: tech SMEs, classic Swiss mid-market, CXO-driven implementation. Other industries on request. We say honestly when a use case is not a fit for us.
Let's get concrete.
30 minutes is enough to find out whether it fits.