Question
What specific data is sent to an external AI provider, where is it stored, and how is confidentiality ensured?
Answer
A regular AI agent in the channel responds to a specific public question from the client — a public comment in a ticket — and uses published knowledge base articles. It does not have access to internal comments and protected confidential data. Client questions and agent responses are saved in the ticket.
When connecting an external model, the conditions for processing and storing queries are determined by the selected AI provider. For processing within the corporate network, you can connect a local OpenAI-compatible model. With on-prem deployment, the main Swarmica data is stored in the client's infrastructure.
If you connect your own external AI agent through MCP, its capabilities and access to data are determined by MCP settings, connection rights, and its skills.
Details: Claim Page: Confidential Data · Connecting a Local LLM · Connection via MCP · Data and Attachment Storage