The accelerated adoption of generative AI across business functions—from finance to engineering—has raised concerns over unauthorised AI usage and inadvertent sharing of sensitive information. Cloudflare’s newly introduced AI Security Posture Management tools aim to provide granular visibility into AI interactions, enforce policy compliance, and safeguard data within a unified platform. Security teams can now identify which AI tools employees are accessing and what data they are sending. The Shadow AI Report delivers detailed insights into usage patterns, moving beyond general detection to reveal specific apps and users.
To counter Shadow AI risks, organisations can apply policy enforcement at network edges using Cloudflare Gateway. This enables the automatic blocking of unapproved AI applications, controlling data uploads, and enabling security-led review processes. At the same time, AI Prompt Protection safeguards against unintentional or risky disclosures by intercepting potentially dangerous prompts—such as those containing source code or customer records—and offering inline warnings or blocking actions before sensitive data is transmitted to external AI services.
Further strengthening control, Cloudflare’s Zero Trust MCP Server Control consolidates outgoing AI model communications into a central dashboard. This allows policy implementation at a user‑level, offering oversight of API usage and better governance over AI tool access.
As part of the broader Secure Access Service Edge architecture, Cloudflare emphasises that its AI‑SPM features—including confidence scoring of AI providers, out‑of‑band API CASB integrations, and visual dashboards—offer a comprehensive control environment that balances innovation and security.
These new capabilities arrive alongside other AI‑focused projects. Cloudflare has opened an Open Beta for MCP Server Portals, accessible via its Zero Trust Dashboard, and launched a user‑research programme to guide future AI security features. Additionally, AI‑driven tools for troubleshooting network and WARP client connectivity issues are now available at no additional cost to Cloudflare One users.
Analysts have responded favourably to these developments. Scotiabank raised Cloudflare’s price target to $200, citing a strong second-quarter performance and expanding role in AI infrastructure—which includes the new AI‑SPM offerings. Meanwhile, Bank of America upgraded Cloudflare stock to “buy”, labelling the company as “one of the true AI winners in software”, thanks to its blend of AI services and security offerings.
Security experts highlight the growing need for zero trust strategies as AI's role in enterprise workflows deepens. At Infosecurity Europe 2025, held in London’s ExCeL Centre, Cloudflare presented its unified security platform amid broader industry discussions on post‑quantum cryptography and AI‑powered threats. Meanwhile, Cloudflare continues to assert leadership across the wider internet infrastructure landscape—its network supports nearly one‑fifth of the web, and its AI ambitions have been characterised by some as turning the company into the "Swiss Army knife" of the internet.
Topics
Technology