> ## Documentation Index > Fetch the complete documentation index at: https://docs.luumen.ai/llms.txt > Use this file to discover all available pages before exploring further. # Integrations overview > How Luumen connects to the external systems your team already operates — ServiceNow, Confluence, observability platforms, and SAP — to enrich monitoring and accelerate response. Luumen integrates with the external systems your team already operates so it can ground its monitoring, scans, and AI in the same context engineers use day-to-day. Four integrations are built in today: **ServiceNow**, **Confluence**, **Observability** (Dynatrace built-in; Datadog, New Relic, and others via white-glove), and **SAP** for environments running SAP landscapes. Other systems can be added during onboarding based on what your environment needs. ## Built-in integrations Read, search, and update SNOW incidents from Luumen. AI-assisted triage that cites past tickets for similar incidents. Pull customer SOPs into Luumen as tribal knowledge. AI uses them to propose Skills and keep documentation current. Enrich host data with telemetry from Dynatrace, Datadog, New Relic, or another provider. Powers advanced AI triage and report views. For SAP environments: collect components, transports, and SAP-specific properties. Powers SAP Security Note matching and transport-aware compliance. ## How integrations work Every integration follows a similar pattern: * **Authenticated connection.** Luumen connects to the external system using credentials you provide during setup (API token, OAuth app, service account). Credentials are stored encrypted at rest and can be rotated through the UI or pulled from [HashiCorp Vault](/enterprise/installation/vault-integration). * **Read-first.** By default, integrations are read-only. Luumen pulls data in (incidents, SOPs, telemetry) and uses it to enrich monitoring and AI responses. * **Explicit write.** Where an integration can write back (e.g., updating a ServiceNow incident with triage notes), every write is associated with a user action and shown in audit logs. There are no silent updates. * **Human-in-the-loop AI.** When LuumenAI uses integration data to propose an action — applying an SOP, updating a ticket, creating a Skill — the action requires explicit approval before it runs. ## Adding integrations The three integrations above are wired into the platform. They're enabled and configured during white-glove setup based on what's in scope for your deployment. Other systems can be integrated as part of onboarding. The pattern is the same: 1. Your team identifies the system and the use case (e.g., "pull alerts from our Splunk into Luumen for cross-referencing with check failures"). 2. Apiphani engineers build or configure the connector and validate it end-to-end in your environment. 3. The integration appears alongside the built-in ones in your workspace. For a new integration, the right path is to raise it with your account team. We'll scope the work and either ship it as part of your existing engagement or add it to the roadmap. See [Support](/enterprise/troubleshooting/support). ## Permissions and audit Integrations honor your existing access control on the external side. The Luumen connector authenticates as a specific user or service account in the integrated system, so anything it reads or writes shows up in that system's audit logs with that identity. We recommend using a dedicated service account for each Luumen integration to make the audit trail unambiguous. On the Luumen side, every integration action is recorded in the workspace audit log alongside the user who initiated it (or the agent run that produced it, for scheduled actions). ## What each integration unlocks | Integration | Read | Write | AI uses it for | | ----------------- | -------------------------------------------------------- | --------------------------------------------- | -------------------------------------------------------------------------------------------------------- | | **ServiceNow** | Incidents, comments, related tickets | Incident updates and comments (with approval) | Incident triage, suggesting resolution paths from similar past incidents | | **Confluence** | Pages in selected spaces (your SOPs and runbooks) | Page edits (with approval) | Skill recommendations, SOP-aware command suggestions, drift detection between SOPs and actual operations | | **Observability** | Per-host metrics and events from your telemetry provider | No (reads only) | Advanced triage grounded in live telemetry, plus report views in the web UI | | **SAP** | System properties, components, transports, BTP data | No (reads only) | SAP Security Note matching, transport-aware compliance checks | Each integration's own page covers setup specifics, capabilities, and limits in detail.