Генерация планов взыскания дебиторской задолженности по просроченным счетам через webhook и HTTP API

## Accounts-Receivable Collection Engine (AI-Powered) Receives an overdue invoice / collection request, enriches customer data, calls an AI service to generate personalized collection actions, executes the primary action, and returns a full collection plan with audit trail. ### All nodes (1-line each) 1. Webhook - Collection Request – Receives the AR collection request via POST. 2. Code - Validate & Build Customer Context – Validates the request and prepares customer risk / history context for AI. 3. Wait - Pre Customer Enrichment – Short pause before customer data enrichment. 4. HTTP Request - Fetch Customer History – Pulls payment history, previous interactions and risk indicators. 5. Wait - Pre AI Generation – Pause before AI action-generation call. 6. HTTP Request - AI Generate Actions – Calls AI service to produce customer-specific collection actions. 7. Wait - Pre Action Execution – Pause before executing the primary recommended action. 8. HTTP Request - Execute Primary Action – Sends email / creates task / triggers call / applies offer. 9. HTTP Request - Update AR / CRM – Logs the action and updates the accounts-receivable / CRM record. 10. Code - Build Collection Plan – Aggregates AI recommendations + execution results into final plan + audit trail. 11. Respond to Webhook – Returns the complete collection plan to the caller.