Citrix added a series of governance updates to its NetScaler platform aimed at securing AI application delivery.

The updates include NetScaler AI Gateway, which governs large language model (LLM) performance, introducing the means to extend existing operational policies and infrastructure controls to AI systems without adding any additional silos.

AI Gateway includes token-based rate limiting, which measures fair consumption across teams and applications; token-latency-based load balancing to optimize performance and reliability by routing inference requests across multiple LLM backends; and spillover routing, which redirects requests to a secondary model when token quotas are exhausted.

AI Gateway, which is part of Citrix's Aidrien AI-powered in-product assistant, acts as a production-ready control layer for AI applications running at enterprise scale.

“AI is becoming embedded in the operational fabric of the enterprise,” NetScaler GM Steve Shah explained. “NetScaler AI Gateway extends our leadership in enterprise application delivery into the AI era, enabling organizations to run AI services with the governance, visibility, and operational confidence required for mission-critical deployments.”

NetScaler also unveiled its Protecto AI-native data protection platform, which adds context security for agentic workloads and pairs with AI Gateway to ensure sensitive information is masked from agents before that data reaches related workflows.

“Protecto's agentic data classification catches what traditional AI security tools miss, and integrates directly with NetScaler AI Gateway to block or mask sensitive data before it ever reaches AI systems – without losing the context AI needs to function accurately,” Protecto Founder and CEO Amar Kanagaraj explained. “Enterprises can finally scale AI at full speed without putting their data at risk.”