AI Assistant Data Privacy
Data Privacy
The Seeq AI Assistant uses enterprise-grade Large Language Models (LLMs) provided by third-party LLM providers. These LLMs are pre-trained on broad datasets and are not trained on Seeq customer private data.
Customer and user data is not used to train LLMs. Seeq does not train LLMs, and our privacy policy, LLM provider configurations, and software controls are designed to prevent customer or user data from being used for model training.
Seeq supports multiple LLM providers and allows customers to opt into a “bring your own LLM provider” (BYOP) configuration. With BYOP, customers use their own LLM deployments, typically via an API gateway such as Azure AI Foundry, and retain full control of the model configuration and inference platform settings.
Your prompts (inputs) and AI Assistant responses (outputs):
Are NOT available to other customers or users.
Are NOT used to train or improve models.
Do NOT cause the models to learn from your usage.
You own your data
You own the prompts you submit and outputs generated for you. You retain all rights to the inputs you provide via the AI Assistant. The models used by the AI Assistant are stateless, which means they do not retain memory from one interaction to the next. Separately, the Seeq AI Assistant securely stores conversation history so users can resume prior conversations from their personal history. A user’s conversation history is accessible to only that user, unless the user chooses to share a conversation with others. Users can delete their own history, or remove conversations shared with them, at any time using the AI Assistant interface.
Security and Compliance
The AI Assistant is included within Seeq’s SOC-2 compliance footprint for Seeq products and services.
Data is segregated and isolated by customer and stored in customer-specific environments.
Data is encrypted at rest using AES-256 and in transit using TLS 1.2 or higher.
Abuse Analysis
Note: this section does NOT apply when using a BYOP configuration
Inputs submitted to LLM providers may be evaluated by automated abuse-detection tools designed to identify prohibited or harmful use of the platform. These tools generate metadata about potential abuse patterns but do not contain any business data.
Inputs and outputs may be securely retained for up to 30 days to detect and investigate platform abuse.
Where is data processed?
Note: this section does NOT apply when using a BYOP configuration
Seeq integrates with multiple LLM providers. Depending on the provider and configuration, inference may be performed on compute resources hosted in different regions globally.