# Use cases Step-by-step scenarios for redacting documents with AI agents over MCP: removing sensitive text by pattern, covering image areas, clearing annotations and metadata, and verifying that a redaction is complete. Step-by-step scenarios for redacting documents with AI agents over MCP — each page answers one real question completely, with the config, the exact prompt, and the expected result: * [How to redact sensitive data from documents with AI agents]({{< ref "redaction/mcp/use-cases/redact-sensitive-data-with-ai-agents.md" >}}) — start here: patterns, chaining, and the four places data hides. * [How to prepare documents for disclosure with an AI agent]({{< ref "redaction/mcp/use-cases/prepare-documents-for-disclosure.md" >}}) — the four-tool pass in order, with the chaining rule that keeps every step. * [How to redact scanned documents and images with an AI agent]({{< ref "redaction/mcp/use-cases/redact-scanned-documents.md" >}}) — no text layer, no pattern matching — coordinates and a solid box. * [How to verify that a redaction is complete]({{< ref "redaction/mcp/use-cases/verify-a-redaction.md" >}}) — re-read, re-search, re-check — a redaction you have not verified is a claim. * [Running GroupDocs MCP servers on-premise: architecture and security model]({{< ref "redaction/mcp/use-cases/on-premise-redaction.md" >}}) — architecture and security model — and what still reaches your model provider. New to the server? [Quick start]({{< ref "redaction/mcp/getting-started/_index.md" >}}) first.