Platform Empowering Enterprises to Securely Deploy Generative AI

Liminal project cover featuring the policy controls for the different supported terms.
Done alongside High Alpha
Client / Liminal
Role / Product Designer
Year / 2023

Liminal provides comprehensive, enterprise-level data protection across various models and applications. Its solutions cater to direct AI interactions, app integrations, and custom-built apps.

Horizontal security platform designed to securely interact with AI models in heavily regulated industries.

Liminal detects non-compliant data (PII, PHI, PCI) in real-time. This safeguarding ensures privacy and compliance without losing the context or intent of the communication, a critical aspect for enterprises handling sensitive client information.

The platform provides users with a secure and direct way to chat with various AI models. It is a controlled environment where users can safely engage with any model that has been enabled by the organization administrators.

Screenshot of the policy control for models

Liminal empowers admins with granular policy controls. Specific rules for different types of sensitive terms can be applied globally, to individual teams, or even to specific users.

Likewise, administrators can manage access to a variety of AI models within Liminal, applying the same detailed level of control. This keeps Liminal adaptable and secure to specific needs.

Auditable logs in Liminal provide comprehensive insights into every AI interaction. They record detailed information alongside the policies applied to each interaction with any enabled model.

Administrators can select and integrate various AI models to the platform. Liminal extends its control to generative AI-enabled commercial apps like Microsoft's Copilot, or Notion.

Screenshot of the Liminal configurations

As the Liminal product evolves, it will bring even more advanced solutions like automated alerts to meet the growing demands of enterprise AI security.