Embodiments of the present disclosure include data loss prevention (DLP) methods and architectures by a cloud-based service. The disclosed techniques of the cloud-based platform (e.g., collaboration platform in an enterprise environment) can detect (and may optionally prevent) violations to, e.g., corporate policies, which can be configurable by a corporate administrator, for example regarding the use, storage, or transmission of sensitive information. The types of sensitive information can include, for example, financial information—credit card and bank account numbers, Personally Identifiable Information (PII)—Social Security Number (SSN), health/healthcare information, Intellectual Property—earnings forecasts, sales pipeline, trade secrets, source code, etc.