Data Processing and Processing Locations
This page describes how and where data is processed within the Evoya platform.
Fundamentals
The decisive level for data processing is the Large Language Model (LLM) assigned to an AI agent.
The selected language model determines where content is processed semantically – regardless of the form in which the data is captured (text, speech, or image).
Processing in Multiple Layers
Data processing occurs in two simplified layers:
1. Preprocessing
Certain content is first technically prepared, for example:
- Voice inputs are converted to text (Speech-to-Text)
- Images are analyzed (Image-to-Text)
- Uploaded documents are prepared for further processing
These steps serve the technical preparation of the content.
2. Further Processing by the Language Model
The semantic processing is then carried out by the Large Language Model assigned to the AI agent. Its location is decisive for further data processing.
Processing Location of the Language Model
The processing location is determined by the selected AI model:
- If a model with location Switzerland is used, further processing takes place in Switzerland
- If a model with location Europe or United States is used, further processing takes place according to that location
The choice of language model is therefore the central factor for data processing.
Image Generation
Separate AI models are used for image generation. The processing location is determined by the selected image generation model and is visible in the model information.
Transparency and Control
The Evoya platform offers full transparency and control over data processing:
- Organizations control which AI models are activated
- AI agents are specifically assigned to particular language models
- The processing location is always traceable