Zammad AI (it@M Integration Layer)

Zammad-AI is a microservice developed by the City of Munich (it@M) that serves as a specialized integration layer for Generative AI in Zammad. Unlike Zammad's native AI features (introduced in version 7.0), this project is designed to implement highly customizable and organization-specific automation workflows. The system consists of three main components: 1. zammad-ai-workflow: The backend service for triage and answer generation. 2. zammad-ai-index: A service for indexing the Zammad knowledge base into Qdrant. 3. slm-guardrails: A service to ensure content safety. By separating these services from the Zammad core, prompt engineering, retrieval logic, and automation rules can be developed and scaled independently of Zammad release cycles. Communication is handled efficiently via webhooks, the Zammad REST API, and Kafka.