Tailscale has introduced new features for Aperture, its AI access and control platform, aimed at providing IT teams with a unified and stable framework for managing AI tools, models, and data sources. This development addresses the growing challenge of ‘shadow AI’—the unsanctioned use of AI applications within organizations.
The proliferation of AI tools in workplaces often occurs without official IT oversight, leading to fragmented and unsecured usage. Employees may utilize personal accounts or adopt various tools independently, resulting in data management blind spots for IT departments. Research indicates that over 64% of activity on personal and free AI accounts is work-related, and companies typically operate nearly 70 generative AI tools, with 90% lacking proper licensing or approval.
Aperture seeks to mitigate these issues by offering a modular AI management solution that avoids vendor lock-in. Its key features include:
- Chat Interface: A browser-based platform enabling employees to access approved AI models, with support for switching between configured large language model (LLM) providers.
- Universal Data Connectors: Tools that facilitate AI integration with internal systems, documents, and operational data, eliminating the need for individual teams to develop their own integrations.
- Identity Preservation: Integration with Tailscale ensures user identities and permissions are maintained throughout the AI agent lifecycle.
- Sandbox Support: Currently in private alpha, this feature provides controlled environments for AI agents to operate without directly interacting with users’ local devices or unmanaged systems.
Aperture is compatible with API keys from major LLM providers, including OpenAI, Anthropic, Google Gemini, and Amazon Bedrock. The chat interface and universal data connectors are now available in public alpha for organizations already using Aperture.
In the rapidly evolving AI landscape, maintaining flexibility and control is crucial for organizations. Aperture’s modular approach allows IT departments to adapt to new tools and vendors without compromising security or oversight, ensuring that AI integration aligns with organizational policies and compliance requirements.
