Identity-based routing, PII guardrails, and secure sandboxing added to Aperture gateway to help IT departments monitor and secure fractured employee AI usage.
Networking startup Tailscale has updated Aperture, its security platform designed to manage and control corporate AI use. The update serves as a unified gateway layer that gives business IT teams direct visibility over how employees and automated systems interact with artificial intelligence models.
This launch addresses the rising issue of unapproved tech tools in the workplace, as recent research shows nearly half of workers accidentally share sensitive company files with public AI software.The updated platform introduces a simple web browser chat interface that lets staff switch between authorized AI models from major providers like OpenAI, Anthropic, Google Gemini, and Amazon Bedrock.
Rather than handing out insecure API developer keys to individuals, the platform automatically logs every request using an employee’s existing corporate network login. This ensures full accountability by mapping every prompt, question, or data request directly to a specific user name while scrubbing out personal identification details before the data ever leaves the corporate network.
For automated software assistants and coding agents, the system introduces a specialized sandbox environment that isolates their activity so they cannot run unmanaged code directly on corporate laptops. Tailscale has partnered with tech monitoring and security companies like Oso, Cerbos, and Cribl to easily log costs, track token counts, and enforce budget caps across different teams. The tool is currently available in a free testing phase for smaller groups, with custom enterprise pricing planned for its wider release.
