
AI is changing how people work across industries: support agents handling many times more tickets, engineers reviewing generated code instead of writing boilerplate, sales teams spending their time on relationships rather than follow-up. Cloudflare says it wants to help companies build AI-driven experiences for employees and customers. Its concern is that the security and governance tooling has not kept pace.
The failure modes are already visible. Vibe-coded apps leak user details; LLM chats meant for colleagues end up indexed by search engines; AI agents get broad access across an organization without proper tracking and control; employees upload confidential company or customer data into an LLM that uses it for training. Content owners face a parallel problem: Cloudflare has found it can be hundreds, or even thousands, of times harder to generate site traffic — and therefore ad revenue — from an AI response than from a search engine result. Leaders are left choosing between blocking all AI usage and letting it run unchecked.
AI Week 2025 is organized around four areas:
- Securing AI environments and workflows — innovation is outpacing control, and Cloudflare positions itself as one of the few zero trust providers that also builds AI infrastructure for the web.
- Protecting original content from misuse by AI — giving creators control over content that AI companies consume as fast as it is produced, with no benefit flowing back.
- Helping developers build secure AI experiences — letting developers run AI applications close to users with security controls built in from day one.
- Making Cloudflare better with AI — applying AI internally, for example to find and mitigate issues buried in logs and events across website, employee and email usage.
Governing AI inside the organization
Traditional security measures are ill-equipped for the risks that come with rapid AI adoption, including Shadow AI and uncontrolled proliferation of models. Cloudflare's argument is that other zero trust providers are not moving fast enough to match AI's pace of innovation. During the week it plans to announce controls to detect Shadow AI and restrict unauthorized AI usage, options for establishing a "paved path" of approved AI tooling that boosts productivity without sacrificing security, and new ways to protect an organization's own models from poisoning or attacks.
Crawl Control and the content question
Cloudflare has introduced Crawl Control to let content owners manage how AI models access and use their content and to address unauthorized scraping and training. Over the past two months it has expanded the number of participating content providers, refined detection so AI crawlers are identified more accurately and only authorized access occurs, and streamlined onboarding so new publishers can begin protecting content within minutes. The stated goal is for creators to be compensated and acknowledged for what they produce.

AI Gateway and developer security
Cloudflare's position is that AI experiences should carry security controls by default, which is why it is investing in both its developer platform's AI Gateway and the associated security controls. Running applications close to end users minimizes latency and improves performance. The week's announcements for AI Gateway include advanced caching to reduce redundant model calls, expanded observability for insight into model performance and usage patterns, and new integrations with popular AI frameworks and services.
AI across Cloudflare's own products
Cloudflare is also embedding AI throughout its product suite, from threat detection that adapts to emerging attack patterns to AI-powered network performance tuning. Announced updates will appear on the AI Week hub page, the Cloudflare blog and CloudflareTV.




