Sherpany AI Security and Ethical Use Policy

Sherpany integrates AI features to enhance productivity and efficiency in meeting management. We do not provide a general-purpose AI platform or train AI models. Instead, we embed verified open-source models from trusted sources into our infrastructure, ensuring full control, data sovereignty, and regulatory alignment.

Sherpany is not an AI vendor

We do not:

❌ Offer a general-purpose AI product

❌  Develop our own AI models

❌  Use customer data to improve or train any model

❌  Provide AI APIs or services to third parties

We only integrate selected trusted open-source LLM models to power specific features within our platform (e.g., document summarisation, rephrasing, translation, transcription).

Privacy and confidentiality

• Local AI deployment: all AI services are deployed and executed entirely within Sherpany's infrastructure, located in Switzerland. No data is sent to third-party AI cloud services or any additional sub-processors.
• Stateless processing: Sherpany AI features are stateless. The AI models do not log, store, or reuse any data used in AI interactions. The Models have no memory of past interactions. When needed, all context is provided with the new prompt.
• No data sharing or exposure: no customer data is used outside our infrastructure. Your documents and prompts are processed in-memory and discarded immediately after.
• Full customer data control: customers remain full owners and controllers of their data. Sherpany retains a processor role, governed by our Data Processing Agreement (DPA), aligned with GDPR.

Alignment with EU AI and Data Protection Principles

• Privacy by design & default: Sherpany’s AI capabilities are built with a “privacy by design” approach: AI features are optional, stateless, transparent, and auditable.
• Explainability and oversight: Sherpany enables human oversight of all AI outputs. Users can review, edit, and validate AI-generated suggestions — which are always optional, never automatic or hidden.
• Data minimisation: we only collect data that is adequate, relevant, and limited to what is necessary for the specific purpose..
• No feedback loops or model drift: because Sherpany’s AI processes are stateless and models are not re-trained, there is no risk of input-based feedback loops or behaviour drift over time.

Intellectual Property and Output Ownership

• Customer data remains confidential: documents processed with AI remain within the secure boundaries of Sherpany's infrastructure and are securely encrypted at rest and in transit. No third party gains access or insight to your documents.
• AI-generated content is non-exclusive and controlled by the customer: any content generated by AI (e.g., summaries or rephrasings) is ephemeral and under the control of the user. Sherpany does not claim ownership of AI-assisted output.

AI at Sherpany: Frequently Asked Questions