The EU AI Act is already in force
Regulation 2024/1689 applies in phases since 2024. Prohibited practices and AI literacy duties already apply, and further obligations follow on a fixed calendar.
ISO 42001 · Liège
In Liège, we build your AI management system to ISO/IEC 42001:2023, from AI inventory to certification audit. You get AI governance that auditors, customers and the EU AI Act can all rely on.

Local context
Space and aerospace firms in Liège often apply AI to imagery, sensor data or the inspection of parts. Their customers include programme primes that expect traceable engineering decisions. A model that flags a defect, or misses one, needs the same rigour as any other quality step. ISO 42001 brings that rigour through lifecycle controls, test records and clear accountability.
University spin-offs in biotech start from a different place. Many are AI providers from day one, building models on clinical or research data. If the product is a medical device, the AI Act treats it as high-risk alongside the MDR. An AI management system set up early makes later conformity work and investor due diligence easier.
Around Liège Airport, cargo and logistics operators use forecasting and screening tools that partners depend on. We work in French or English, on site in the Liège area, including the airport and science park zones. The city is about one hour from our Brussels office by train, which suits recurring on-site workshops.
Liège combines Liège Airport, a major European cargo hub, with a space and aerospace cluster, university spin-offs in biotech and a long industrial tradition. These organisations handle sensitive data, regulated products and operational technology that customers and authorities expect to be secured.
Liège is about one hour from our Brussels office by train. We work on site in the Liège area, including the airport and science park zones, or remotely.
What is at stake
Regulation 2024/1689 applies in phases since 2024. Prohibited practices and AI literacy duties already apply, and further obligations follow on a fixed calendar.
Procurement teams now send AI questionnaires next to security ones. Without a documented system, each answer is improvised and deals slow down.
Bias, opacity, model drift and misuse are not covered by a classic security risk register. They need their own assessment and owners.
Key sectors in Liège
Route planning and demand forecasting models drive daily operations, and ISO 42001 makes their monitoring and fallback explicit.
AI in design or inspection must meet strict traceability expectations, which the AI lifecycle controls of ISO 42001 support.
Clinical and administrative AI handles sensitive data and affects patients, so impact assessment and human oversight are essential.
Predictive maintenance and quality inspection models need data quality rules and human review to stay trustworthy.
What we do
We do not start everything at once. The scoping call shows which workstream matters most, and we start there.
We list every AI system you build, buy or embed, including third-party APIs. We confirm your role for each: provider, deployer or user.
We extend your risk method to AI-specific harms. We run the AI system impact assessment required by the standard for each in-scope system.
We write the AI policy, roles and lifecycle procedures from design to retirement. Controls cover data quality, transparency, human oversight and suppliers.
We run the internal audit and the management review, then prepare your teams for the certification body. If you hold ISO 27001, we extend that system rather than duplicate it.
A 30-minute call is enough to scope your AI management system and its next steps.
Regulators
Centre for Cybersecurity Belgium: national NIS 2 authority, registration and incident notification.
Belgian Data Protection Authority: GDPR supervision and complaints.
Financial supervisors for banks, insurers and investment firms, including DORA.
Belgian law of 26 April 2024, in force since 18 October 2024, with CyberFundamentals as the reference framework.
The method
Each step ends with a document you keep, even if the engagement stops.
A free 30-minute call to understand your AI use, your customers' demands and your timeline.
Deliverable : Written scope and fixed price per lotWe interview product, data and IT teams and list every AI system in scope.
Deliverable : AI system register with roles and ownersWe compare current practice with clauses 4 to 10 and Annex A of ISO/IEC 42001.
Deliverable : Gap analysis report with prioritised actionsWe assess each system for risks to the organisation, to individuals and to society.
Deliverable : AI risk register and impact assessmentsWe draft policies, procedures and the Statement of Applicability, then help teams apply them.
Deliverable : AI management system documentation setWe audit the system, run the management review and support you during the external audit.
Deliverable : Internal audit report and certification readiness fileScope and budget
Questions
Not if it is scoped well. The standard scales to the AI systems you actually run, and a small team can own it. We start with the inventory, the impact assessment and supplier controls, then add what certification requires.
Yes. These standards follow a similar management system logic: context, leadership, planning, operation, evaluation and improvement. We map ISO 42001 onto your existing quality processes so audits and records stay unified.
A large part. Both standards share the same management system structure. Document control, internal audit and management review can be extended rather than rebuilt.
Yes. We work on-site or remotely, in French, Dutch or English. Workshops with product and data teams often work best in person.
No. Certification is issued by an accredited certification body, which must stay independent from the consultant. We prepare you and support you during the audit.
Last step
Tell us how you use AI today. You receive a written scope and a fixed price per lot after our call.