The flow
Starts it 2Model 2Logic 11Service 4
The steps in the order they run, from what starts it.
- 1Schedule Trigger - Poll Complaint SourceSchedule triggerStarts it
- 2Manual Trigger - Test RunManual triggerStarts it
- 3Anthropic - Complaint Extraction ModelLm Chat AnthropicModel
- 4Set - ConfigSetLogic
- 5AI Agent - Extract Structured Complaint FieldsAgentModel
- 6HTTP - Fetch New ComplaintsHttp RequestService
- 7JS - Parse Extracted FieldsCodeLogic
- 8JS - Normalize Complaint RecordsCodeLogic
- 9JS - Cluster by Product & Defect CategoryCodeLogic
- 10IF - New Complaints Found?IfLogic
- 11JS - Compare Clusters Against Rolling BaselineCodeLogic
- 12JS - Fan Out Complaints for ExtractionCodeLogic
- 13No Action - End CycleNo OpLogic
- 14JS - Fan Out Clusters for RoutingCodeLogic
- 15Switch - Route by Statistical AnomalySwitchLogic
- 16HTTP - Create CAPA Quality EscalationHttp RequestService
- 17Merge - Combine Escalation & Routine PathsMergeLogic
- 18Slack - Alert Quality Team (Cluster Escalation)SlackService
- 19HTTP - Log Cluster to Trend StoreHttp RequestService