Cloudflare Launches Identity-Aware AI Gateway to Track Who Is Using AI and Control Costs

Cloudflare Tackles Enterprise AI Visibility Gap

Cloudflare launched Identity-Aware AI Gateway, addressing two critical challenges facing enterprises: spiraling AI costs and security blind spots1

1

2

2

. IT teams can now track who is using AI by attaching a verified identity to every AI request leaving company networks. The service integrates with existing infrastructure, requiring no new systems while delivering visibility and control over employee AI usage that was previously impossible3

3

.

Linking Identity to Every AI Request

The Identity-Aware AI Gateway connects directly to Cloudflare Access, the company's zero-trust access product, eliminating blind shared API keys that make attribution impossible1

1

. Companies integrate with their Identity Provider and Zero Trust network access infrastructure to validate user and device posture before requests reach AI models. Verified identities are written directly into request logs, giving security teams precise accountability2

2

. Max Baumgarten, a security engineer at Flexport, confirmed that shared API keys made it "almost impossible to tell who is using an AI service" before implementing the solution1

1

.

User Insights Detects Anomalies and Waste

The companion User Insights feature learns baseline AI usage patterns for every person and automated system on a network2

2

3

3

. When spending spikes beyond normal levels for a specific user or system, IT teams receive immediate alerts. The feature also identifies efficiency opportunities by checking whether employees are pushing simple tasks through expensive, high-powered AI models when lighter alternatives could handle the workload. This behavior anomaly detection addresses a growing problem: while individual AI requests keep getting cheaper, employees and bots fire off far more of them, causing bills to spike with little warning1

1

.

Cost Control and Security Safeguards

IT teams can set spending limits at individual or team levels, preventing runaway usage before it hits invoices2

2

. Rate limiting stops excessive requests in real-time, while automatic filters strip employee names, passwords, and other sensitive data from requests before they reach outside AI providers1

1

3

3

. The system also caches repeat requests to cut redundant costs. These capabilities address security risks where confidential files can pass through to external AI models with nobody noticing, a concern that has pushed companies to lock down AI access entirely.

Shifting from Restriction to Enablement

Dane Knecht, Cloudflare's Chief Technology Officer, framed the launch as a fundamental shift in enterprise AI strategy. "When security is clunky, it becomes a speed bump that slows down innovation," Knecht explained2

2

. Fear of surprise bills and accidental data leaks has driven companies to restrict what teams can build. "Connecting our access controls directly to AI Gateway flips that dynamic," he said. "Teams get the freedom to work with any AI model they choose. IT gets real-time visibility, guardrails, and the budget controls they actually need"1

1

. The AI Gateway itself dates back to September 2023, with security controls added to Cloudflare's zero-trust platform since August last year1

1

. The new identity-aware capabilities represent the latest evolution in helping enterprises balance innovation speed with governance requirements as AI usage continues its rapid expansion across organizations.