AI security and IP protection in Switzerland
SteP-met combines context-based AI with a privacy-focused architecture and controlled data processing. Relevant and authorised information, clear permissions and the responsible handling of each respective context are at the heart of this approach. At the same time, all confidential know-how, intellectual property, and sensitive implementation details are securely protected.
Privacy-focused architecture for context-based AI
SteP-met’s approach starts with a clear question: What information is actually needed for any specific usage strategy? The AI should never access data indiscriminately. Instead, relevant, authorised information from each respective learning, performance, support or business context must always be provided in a targeted manner.
The SteP-met architecture is thus designed for controlled data processing. Every specific implementation is based on the data protection, security and governance requirements of the respective company or institution. SteP-met deliberately makes no blanket promise to comply with every conceivable data protection regulation. In practice, such requirements are carefully considered and implemented in line with the project-specific demands.
Pseudonymisation and the responsible handling of identifying data
Not every AI-supported analysis requires real names. Where a specific usage strategy allows, SteP-met relies on pseudonymised processing or avoids using real names in AI contexts.
This principle is particularly relevant when learning, training, performance or other personal context information is processed. The key is to provide just the information required and authorised for the respective function.
This means data processing always follows SteP-met’s fundamental principle: Data & systems → secure integration → relevant context → AI → practical support.
Authorised data instead of disconnected AI
At SteP-met, context-based AI is deliberately different from an all-inclusive generative application. Analyses, recommendations and support functions are thus based on information that is both relevant and authorised for the predefined usage strategy.
The actual data provided as context thus depends on the respective solution and its applicable permissions. This could be learning progress in an educational context, defined performance information in sport, or a relevant body of knowledge for business support.
This approach creates a clear link between the data source, the authorisation protocols, and the intended function. This clarifies precisely which context is intended for AI support.
Strategic protection for know-how and intellectual property
SteP-met deliberately separates the public overview from any confidential technical implementation details. Interested companies and institutions receive a comprehensible overview of the basic principle on the SteP-met website without the disclosure of protected technical know-how.
Clarify your security and governance requirements at an early stage