Yes and no. Unlike free tools (e.g., ChatGPT), your data never trains models and is protected by bank-grade encryption. Your client documents stay yours.

However, without a Custom Workbench, your data is at risk of being trained on by the LLM provider. If you decide to use a Custom Workbench, you can optionally add a private LLM which is 100% secure and dedicated only to you and your organization.

Ultimately, the only way to be completely secure with AI is to use a trusted LLM provider (like Mill Pond Research) or privately host your own LLM.