Open information about our AI systems according to EU Regulation
Pursuant to Article 50 of Regulation (EU) 2024/1689 (EU AI Act), we are required to provide transparent information about the use of AI in our service. This page provides comprehensive information about our AI systems, their capabilities, limitations, and your rights.
Classified as limited-risk AI system per Annex III
Our system fulfills all transparency obligations pursuant to Article 50 EU AI Act
We exclusively use AI models from Mistral AI (France) - all data processing occurs within the EU:
Analyzes your notice documents for errors, inconsistencies and potential claims under SGB II. EU-hosted (France).
Creates vector embeddings for searching legal documents and laws (RAG system). EU-hosted (France).
Extracts text from uploaded PDF files and images for further analysis. EU-hosted (mokka-dev.de).
You upload your citizen's benefit notice as PDF or image
Text is extracted and analyzed by AI for errors
A specialist attorney reviews results and advises you further
Automatic text recognition (OCR) from PDFs and images
Identification of typical errors in citizen's benefit notices
Calculation of costs and benefits
Detection of possible claims for higher benefits
The AI analysis does not replace professional legal advice from a specialist attorney
The system may have human errors in data processing
Not all types of errors are detected – manual review is recommended
Your documents are processed securely and not used for AI model training.
All AI analysis results are reviewed by specialized attorneys in social law before being implemented in legal steps. This ensures that potential errors are detected and corrected.
You can request a comprehensible explanation of the AI decision-making logic
Your case is automatically reviewed by a specialist attorney
In case of violations of EU AI Act you can complain to the supervisory authority
You can request deletion of your data at any time
Pursuant to Article 11 EU AI Act, we maintain comprehensive technical documentation of our AI system. This documentation contains details about system architecture, data processing, risk assessment, and security measures.
Request DocumentationFor questions or complaints regarding our AI system, you can contact the competent supervisory authority:
Bundesnetzagentur - Competent market surveillance authority for AI systems
