The document-generation side of the AI surface, sold as a separate module from the in-workflow AI Assistant.
Where AI Assistant lives as sparkle buttons inside forms (risk suggestions, gap analysis, next-best-action), AI Documents owns the long-form output: policies, procedures, risk assessments, business continuity plans, ISO 27001 SoA — generated from your instance data (assets, business processes, risks, controls, responsible persons) plus a structured questionnaire, with the AI drafting the connective prose.
- Template-driven — every document type has its own template and questionnaire schema
- Data-aware — pulls assets, risks, BPs, controls already in your instance so you don't restate them
- Lifecycle aware — generated drafts flow through the 5-stage document lifecycle (draft → review → approved → signed → archived) like any hand-written document
- Control auto-linking — when a generated document is approved, the controls it satisfies are linked automatically via the template's glob patterns
- Standards-aware — generation is conditioned on the standards active in the instance, so the output matches what the auditor will look for
Three provider modes — same plumbing as AI Assistant and Ludwig. Bring your own key (Anthropic, OpenAI, Gemini, Azure OpenAI), bring your own self-hosted endpoint (Ollama, vLLM, in-house LLM for air-gapped deployments), or let RaulWalter supply the AI key for a one-invoice arrangement. Off by default until the administrator turns it on.
Bought separately from AI Assistant (in-workflow sparkle buttons) and Ludwig (standalone chat advisor). The three AI modules combine for organisations that want full coverage; many customers take just one or two.