Assess Your EU AI Act Compliance

Critical milestones 2025-2027

February 2025 - Prohibition of prohibited AI systems

- Obligation: Mandatory audit of existing systems

- Sanction: Up to €35M or 7% of global turnover

- Example: An SME using an automated HR scoring system without human supervision must cease immediately or risk maximum sanctions.

August 2025 - GPAI obligations (General Purpose AI)

- Obligation: Transparency documentation for generative AI

- Sanction: Up to €15M or 3% of global turnover

- Example: An SME deploying ChatGPT Enterprise must document usage, risks and mitigation measures under penalty of fine.

August 2026 - Full compliance for high-risk systems

- Obligation: Compliance assessment + mandatory certification

- Sanction: Market ban

- Example: An industrial company using predictive maintenance AI on critical infrastructures must obtain CE marking before operation.

August 2027 - Full application of all provisions

- Obligation: Continuous monitoring + annual audits

- Sanction: Market withdrawal + cumulative fines

- Impact: Permanent surveillance of all AI systems with mandatory reporting.

APOLLO Framework™ AI Act Methodology

Developed by consultants certified MIT Sloan Executive Program "Artificial Intelligence: Implications for Business Strategy".

Dual scoring: General 0-100 classification + Sector-specific compliance thresholds

7 business modules: Management, IT, HR, Sales, Procurement, Operations, Legal

Official EU standards: Calibrated on Regulation (EU) 2024/1689 and Commission guidelines

Personalized action plan: Prioritized recommendations with regulatory justifications

Sector Use Cases

🏭 Industry & Manufacturing

Use Case: High-risk (Annex III - critical infrastructure)

AI Act risk level: 42/100 - Non-compliant

Average SME compliance score: 42/100 - Non-compliant

Priority actions: Impact assessment + Technical documentation + Robustness testing

Concrete example: An industrial SME uses predictive AI on production lines. Assessment result: Score 38/100, "Non-compliant" status. Urgent actions: Conduct fundamental rights impact analysis (Article 27), document technical architecture, implement continuous performance monitoring.

🏦 Banking & Insurance

Use case: Fraud detection + alternative credit scoring

AI Act risk level: High-risk (Annex III - access to essential services)

Average SME compliance score: 56/100 - Partially compliant

Priority actions: Decision explainability + Human review rights + Treatment registry

Concrete example: A regional bank deploys alternative AI credit scoring. Assessment reveals: Score 53/100, critical gaps in explainability and human oversight. Recommendations: Implement LIME/SHAP explainability interface, guarantee 100% human review of rejections, document GDPR-compliant registry.

🛒 Retail & Distribution

Use case: Personalization + intelligent inventory management

AI Act risk level: Limited risk (transparency required)

Average SME compliance score: 68/100 - Acceptable

Priority actions: User information + AI cookie policy + Opt-out

Concrete example: An e-commerce platform uses AI product recommendations. Assessment: Score 65/100, "Acceptable" status but improvements needed. Actions: Add visible "AI-recommended" labels, implement personalization opt-out, audit bias across customer segments.

💊 Healthcare & Pharma

Use case: Assisted diagnostics + patient journey optimization

AI Act risk level: High-risk (Annex III - medical devices)

Average SME compliance score: 35/100 - Non-compliant

Priority actions: CE marking certification + Clinical validation + Decision traceability

Concrete example: A healthcare facility tests AI radiology diagnosis. Critical assessment: Score 32/100, operation forbidden without compliance. Urgent obligations: Obtain medical CE marking, conduct validated clinical trials, guarantee 100% decision traceability with physician override.

The Diagnostic in Action

Intuitive interface

1. Identify your profile: Sector (5 categories) + Company size

2. Map your AI systems: 7 business modules evaluated

3. Answer 63 questions: Duration 45-60 minutes

4. Risk classification: Automatic per AI Act Annex III

Instant results

• Global score: 0-100 with compliance level (5 categories)

• Module analysis: 7 detailed scores + compliance statuses

• Action plan: Prioritized recommendations

• Professional PDF report: 6 pages with glossary + regulatory authorities

Specialized FAQ

How much does non-compliance cost ?

Sanctions range from €15M to €35M or 3% to 7% of global turnover. For an SME with €50M turnover, this represents up to €3.5M maximum fine.

Who in my company should participate ?

Heads of each department concerned by AI: Management, IT, HR, Commercial, Legal, Production, Procurement. The tool guides the distribution of questions by business module.

What is the validity of my assessment ?

We recommend quarterly reassessment because: (1) The AI Act regulation evolves with new guidelines, (2) Your AI systems evolve, (3) Sectoral authorities publish clarifications. An assessment older than 6+ months may be obsolete.

Does this assessment replace a legal audit ?

No. This tool provides a structured diagnosis and actionable recommendations but does not constitute legal advice. For specific regulatory interpretation, consult a specialized AI law firm.

What to do if I am non-compliant ?

The report provides a prioritized action plan. We recommend starting with "urgent" actions, then structuring a compliance plan over 12-18 months.

What are your criteria based on ?

Our methodology integrates the official European Regulation, European Commission guidelines, CNIL and NIST standards, adjusted to sectoral specificities with specialized legal firm validation.

"The assessment revealed 3 unidentified high-risk AI systems. The detailed report with precise regulatory references enabled us to structure our compliance plan and convince the board of a €180k budget."

Laurent M., CIO, Automotive Industry
(350 employees)

★★★★★

"In 1 hour, we obtained comprehensive mapping of our AI Act compliance gaps. The prioritized recommendations with regulatory justifications are directly actionable. We shared the report with our legal firm who validated its relevance."

Sophie D., Compliance Director, Fintech
(120 employees)

★★★★★

Anticipate AI Act obligations 2025-2027

Companies structuring compliance today avoid tomorrow's penalties. First deadline: February 2025.

Secured confidential data, Professional PDF report included, Immediate results