Cloudflare Confidence Scorecards - making AI safer for the Internet

Cloudflare has launched Application Confidence Scorecards as part of its AI Security features within the Cloudflare One SASE platform to help manage the risks of Shadow AI. The scores assess AI-enabled and SaaS applications across two rubrics: the Application Posture Score and the Gen-AI Posture Score, evaluating security, privacy, data management, deployment security, and training data governance, among other factors.

Scores are produced automatically by a system that crawls public sources, with large language models parsing documents and human analysts validating the results before they appear in the Application Library. The intention is to give decision makers a clear, live view of each tool’s safety and trustworthiness, so policies can be set or guardrails applied accordingly.

Looking ahead, Cloudflare plans to use these confidence scores to drive enforcement within Cloudflare One, enabling gateways to block or warn about low-scoring apps and to link DLP policies to confidence scores. The features are now live in the Application Library and available to users, including those with free accounts. 23 September 2025.