Our method
AI is not a tool, it is an architectural decision.
Our method makes that decision verifiable — in four stages, with clear stopping points. It runs the same way in every engagement, whatever the sector, country or regulation.
Why structure comes before technology
AI amplifies what is already there.
Which is why the organisation decides the outcome, not the model.
That is not a stance, it is the finding from every audit we have run so far.
Good structures + AI
dramatic acceleration
Chaos + AI
faster, larger chaos
What AI cannot do
- Repair broken processes
- Dissolve organisational chaos
- Eliminate technical debt
What AI makes visible
- Inefficiencies in the workflow
- Data quality problems
- Process weaknesses nobody had named before
The most common data problem is not one of quality or volume. It is one of access.
The three principles
Organisation before technology.
We do not ask which model. We ask which process, which decision, which accountability. That rules out every recommendation that starts with a product choice.
People before algorithms.
A system nobody operates is not a system. That rules out every architecture designed without the people who will later work with it.
Impact before scale.
A use case has to prove its value in operations before it can scale. That rules out a rollout before the results are proven.
Three working principles
Not mission statements but guard rails. They restrict what we are allowed to recommend in the first place.
- Problem first, AI second
- Every initiative addresses a named business pain — never a technology.
- Agentic first
- Build data, systems and processes from day one so that agents can read, write and operate them.
- Enablement over dependency
- The capability stays in your organisation, beyond the contract. Dependency is not a business model.
The four stages
Four stages with gradually less involvement from us.
At the end the organisation runs without us. Our involvement tapers by design — from the outset, as part of the offer rather than as a statement of intent. Every stage has exit criteria that can be checked rather than interpreted.
Audit
Entry pointStatus and potential. Always first. For some questions it is already the complete answer.
Decision brief delivered to the management team
Foundation
90 daysBuild foundations and governance, implement initial use cases. These usually go live during the Foundation, or later depending on the starting point and available IT capacity. We agree the timeline and milestones together and review them every three months.
Tools ready for use, AI lab established, first quick wins in implementation
Scaling
Month 3–9More use cases, training, building internal AI champions. From month 6 the focus shifts to handover.
Internal AI leadership able to act, champions active, handover underway
Anchoring
Month 9–18The organisation develops further on its own. We remain available for specialist questions and strategic discussions, with gradually less involvement.
Internal teams operate, improve and extend AI applications independently
Reducing our involvement is conditional. It depends on internal roles being filled — not on our calendar.
The audit
How we conduct an audit.
We show the method because the method is the proof, not the claim. The audit is the entry point to every engagement and can also be booked on its own.
Sequence
A review of the organisation, the system landscape and initiatives already running. The output is a hypothesis-driven workshop framework. We arrive with hypotheses, not with a questionnaire.
Readiness
Four dimensions, scored 1 to 5.
Four dimensions, one overall picture. Move the sliders to see where your organisation stands and which development step comes next.
AI Readiness Score
2.5/ 5
Systematic development
The foundations are taking shape. The next step is to turn successful applications into repeatable processes and dependable operations.
Next development step
- Strengthen your data foundation: Clarify ownership, data quality and reliable access.
- Enable agent access: Provide suitable interfaces for reading and writing with clearly defined permissions.
Current bottlenecks
- Data architecture & data quality2 / 5
- Software stack & agentic interoperability2 / 5
In the audit, this assessment is verified and supported by interviews and a review of your systems.
Use-case portfolio
The right use cases. In the right order.
- Derived from your interviews, assessed by business value, total cost and payback.
- From €10,000 in annual value; implementation follows prerequisites and dependencies.
- Cross-functional use cases benefit multiple departments: a shared technical foundation reduces duplicated development and operating costs.
We identify organisational prerequisites separately.
We make gaps in the data explicit. We calculate returns from go-live.
Your outcome
- A readiness score with supporting rationale and a prioritised portfolio with a business case.
- A 12–18-month roadmap with phase targets, dependencies and responsibilities — a basis for management decisions.
Sample document available on request.
When this pays off
Making investment and returns transparent.
First, the foundations
- Data integration, governance, training and initial applications require upfront investment. An early go-live does not mean the total investment has paid for itself. We therefore assess returns for individual use cases and the overall programme separately.
Then, share and reuse
- Data connections, approval processes and AI capabilities become available to other departments. A central skill repository makes solutions reusable and keeps knowledge within the business.
The economic benefit grows when multiple use cases share the same foundation. We calculate when the investment pays off based on your starting point.
What we anchor in the organisation
So your organisation can govern AI independently.
Internal AI responsibility
An empowered role connects technical competence with organisational governance.
Champions in each department
Internal contacts support colleagues and share experience.
Clear decision paths
Who decides what, by when — and where does an approval go if it stalls?
Reliable use of AI tools
Approved tools and clear rules make uncontrolled use less attractive.
Appropriate human oversight
Reviews and approvals reflect the risk and impact of each application.
AI governance and the EU AI Act
We consider the EU AI Act requirements relevant to your AI applications and work with your responsible teams to translate them into responsibilities, documentation and controls.
What international companies also need to consider
- We consider which models and platforms are available, permissible and operationally usable locally. This creates appropriate solutions for each location — with shared standards and clear responsibility between headquarters and local teams.
Next step
Every route begins with the audit.
In 30 minutes, we get to know each other and offer an initial perspective on your situation. Together, we explore whether working together and an audit as a next step would be a good fit. Not a sales call.