Cloudflare, the cloud connectivity company that protects millions of organizations worldwide, has announced the launch of a set of capabilities designed to secure the adoption of generative artificial intelligence (AI) within companies. This new integration with Cloudflare One, its Zero Trust security platform, aims to address an increasing challenge: how to deploy AI applications at scale without risking privacy, corporate data, or system stability.
The announcement comes at a time when AI tool usage has surged across various departments—from marketing and finance to engineering and design—often without security teams’ oversight. The result: potential leaks of sensitive information, manipulated models through prompt injection attacks, and deployments without proper controls.
“Cloudflare is the best place to help any business deploy AI securely. We are the only company that combines a Zero Trust platform with a comprehensive suite of AI and inference products, all backed by the scale of a global network,” said Matthew Prince, CEO and co-founder of Cloudflare.
A clear picture of AI use in businesses today
The first major innovation is called AI Security Posture Management (AI-SPM). With this feature, organizations can discover how employees are actually using AI applications. Thanks to the new Shadow AI Report, security teams will gain granular insights—detailing not just whether someone is using a chatbot, but which one specifically, and what data is being shared.
This approach responds to a phenomenon known as Shadow AI, which involves uncontrolled use of AI applications by employees, often sharing sensitive company data.
Protecting against data leaks and misuse
Cloudflare has incorporated controls to detect and block risky prompt interactions. The feature, called AI Prompt Protection, analyzes user requests to AI models and can warn or block the transmission of confidential information, such as source code or customer data.
This allows the company to strike a balance between security and productivity: not banning AI in the workplace, but enabling its safe use without exposing trade secrets or violating privacy regulations.
Control over AI interactions with external systems
Another important development is Zero Trust MCP Server Control, a dashboard that centralizes all calls made by AI models to external servers (the so-called MCP tools). This provides security teams with the ability to establish access policies and control at both user and server levels, with full traceability.
Beyond security: productive confidence
This announcement reinforces Cloudflare’s vision of becoming the key partner for companies wanting to harness AI opportunities without succumbing to risks. The company reminds that it blocks billions of threats daily on its global network, and its platform is already deployed in some of the world’s largest organizations, as well as startups, NGOs, and governments.
With these new features, Cloudflare aims to give clients confidence that AI projects—ranging from internal chatbots to multimodal applications—can be deployed securely from day one.
A global challenge: security in the AI era
This initiative also reflects the increasing pressure on companies in a landscape where AI is becoming more pervasive. Incidents of prompt injection, leakage of trade secrets through AI tools, and security-uncertified deployments have made headlines in recent months.
Cloudflare’s clear promise: provide a framework that allows teams to innovate without fear of risks paralyzing their projects.
Frequently Asked Questions (FAQ)
What is Shadow AI and why is it problematic?
Shadow AI refers to the use of AI tools within an organization without security teams’ knowledge or oversight. This can pose risks such as leaks of confidential data, regulatory violations, or exposure to model manipulation attacks.
What is Cloudflare’s AI Prompt Protection?
It is a feature that reviews prompts and responses from employees with AI models in real-time. It can warn or block the submission of sensitive data—like proprietary code or client information—before they are sent to external services, reducing the risk of data leaks.
What does Zero Trust MCP Server Control offer?
It centralizes and audits all calls made by AI models to external servers and tools, applying security policies at both user and infrastructure levels. This provides greater control and visibility over AI system integrations.
Why does Cloudflare believe it has a security competitive advantage for AI?
Because it is the only company offering an integrated Zero Trust platform combined with a comprehensive ecosystem of AI inference and development products, all supported by a global network processing billions of requests daily and blocking threats in real time.
via: cloudflare