AI Transparency

1. How we use AI

At Monitoris, AI is used in particular to support:

• the processing and analysis of documents such as receipts and invoices
• the structured capture of information
• assistance within the digital back office
• the automation of recurring administrative tasks

2. AI technologies used

Monitoris uses modern AI technologies, including:

• large language models (LLMs) for text and document processing
• generative AI components for the structured preparation of information
• depending on the function, additional technologies such as OCR or speech-to-text systems

3. Origin and operation of the models

Depending on the use case, Monitoris uses: established commercial models from external providers, or selected open-source models The operation takes place on secure infrastructures. Hosting may be located in Switzerland, the European Union, or, depending on the service used, in third countries. In all cases, we ensure compliance with data protection and security requirements through appropriate contractual and technical safeguards.

4. No training with customer data

A core principle at Monitoris is: We do not train AI models with customer data. User inputs, documents, and content are processed solely for the purpose of providing the respective functionality and are not used for model training or further development unless explicitly stated otherwise.

5. Control instead of a black box

We do not adapt the behavior of AI systems through uncontrolled training. Instead, we rely on:

• clear system and usage guidelines
• structured prompts and rules
• context-aware processing
• technical protection mechanisms

6. Responsibility and limitations

AI systems can make mistakes. Therefore:

• AI-generated results do not replace professional review
• users remain responsible for decisions
• critical processes are deliberately limited or safeguarded