EU AI Act · Technical Compliance
EU AI Act.
Your obligation.
Our implementation.
The AI Act is in force. Companies that integrate LLMs, use automated decision-making or deploy AI-assisted systems need to act now — before the fines become real.
Max. fine (prohibitions)
35 Mio. €
Max. fine (obligations)
15 Mio. €
Or % of global revenue
7 %
Transparency deadline
Aug 2025
Risk pyramid
Where does your AI system sit?
The EU AI Act divides AI systems into 4 risk levels — each with different obligations and timelines.
Unacceptable risk
PROHIBITEDMust be immediately discontinued
High risk
STRICT REQUIREMENTSFull technical documentation + conformity assessment required
Limited risk
TRANSPARENCYMandatory disclosure to users that they are interacting with AI
Minimal risk
VOLUNTARY CODEVoluntary codes of conduct recommended
Our deliverables
From assessment to production.
Risk Classification
Systematic assessment of all your AI systems according to EU AI Act categories. Result: binding classification with legal basis and recommended action plan.
Technical Documentation
Complete technical documentation according to Annex IV EU AI Act: system description, training data, architecture, performance metrics, risk management.
Audit Trail & Logging
Automatic logging of all relevant AI decisions, inputs and outputs. Immutable, GDPR-compliant log infrastructure with retention policy and access control.
Monitoring Dashboard
Live dashboard for model performance, bias detection and anomaly alerts. Automated reports for internal governance and possible authority inspections.
Human Oversight
Implementation of the mandatory human review mechanisms: override functions, escalation protocols, "meaningful human control" according to Art. 14 EU AI Act.
Transparency Layer
UI/UX components that inform users when AI is involved in a decision. Disclosure texts, opt-out mechanisms, explainability views — GDPR Art. 22 compliant.
Enforcement timeline
You have less time than you think.
EU AI Act entered into force
The Regulation is part of EU law. 24-month clock starts for most obligations.
Prohibitions apply
All unacceptable-risk AI systems must be discontinued immediately. No exceptions.
Transparency obligations
All AI systems must inform users that they are interacting with AI. Chatbots, recommendation systems, content generators — all affected.
High-risk AI: full compliance
Complete technical documentation, conformity assessment, audit trail and human oversight for all high-risk systems.
GPAI models & full enforcement
General-purpose AI models (GPT-4 class) with additional transparency obligations. Full EU-wide enforcement with supervisory authorities.
Pricing
From assessment to full compliance.
Assessment
Can be credited against implementation
2-day compliance audit of all your AI systems. Result: classification, gap analysis, prioritised action plan.
- All AI systems inventoried
- Risk classification per system
- Gap analysis vs. AI Act
- Written action plan
- Q&A session (2h)
Compliance Package
Full implementation for 1–2 medium-risk AI systems: documentation, audit trail, transparency layer.
- Everything in Assessment
- Technical documentation (Annex IV)
- Audit trail implementation
- Transparency UI components
- Human oversight mechanisms
- 3 months post-go-live support
High-Risk
Conformity assessment, notified body coordination, complete documentation for high-risk AI systems.
- Everything in Compliance
- Conformity assessment procedure
- Notified body coordination
- Ongoing monitoring dashboard
- Annual compliance audit
- Regulatory legal support
Technical stack for AI compliance
PostgreSQL Audit LogGrafana
Next.jsPython FastAPIZod Schema ValidationBundle with AI integration
Build compliant — from day one.
If you are planning a new AI integration, we combine technical implementation and legal compliance in one project. That saves 30–40 % compared to retrofitting compliance into an existing system.
