No external models. Architectural, not a setting.

NexDiscovery does not transmit customer data to any external AI model under any circumstance. No OpenAI, Anthropic, Google, or Azure OpenAI. All analysis runs inside the customer's own infrastructure.

This is an architectural constraint, not a configuration setting. There is no token cost because no external inference engine is used.

Every finding the platform surfaces is reviewed by your team before any action, and each one references the source records that produced it.