Features
- Bulk import engine - New processing engine with improved validation, mapping, and error handling for importing large datasets. Includes support for flat-shaped CSV files and row-by-row upsert operations.
- Enhanced AI agent framework - Added specialized agents for master data operations, close management, bulk operations, and fixed assets with improved tool delegation capabilities.
- Avalara tax integration - Enhanced push transaction logic with improved customer credit mapping and refined default integration settings for better tax compliance workflows.
- Slack user lookup - New capability to search and resolve Slack users by name or email within integrated workflows.
- Record source line API - New unified endpoint for fetching records from multiple sources with client-side sorting, filtering, and pagination support.
- Anomaly detection filters - Added support for multiple transaction type filters with improved selection retention in anomaly detection workflows.
- Ramp integration settings - Added configuration option to use approval date as reimbursement posting date for more accurate transaction timing.
- Entity mutation events - Simplified entity update notifications with direct event emission, replacing the previous artifact-based system.
- Embedded Copilot context - Enhanced AI copilot with surface context awareness, allowing the assistant to understand what page or entity the user is currently viewing.
Fixes
- Bank Match archival - Fixed issue where bank matches on termination transactions prevented proper archival of contracts.
- Brex integration sync - Resolved sync issues and extra GL account creation problems. Fixed account deactivation in Brex when accounts are archived in DualEntry.
- Ramp transaction sync - Fixed timing issues where Ramp transactions were not syncing simultaneously, ensuring consistent data flow.
- Change order credits - Corrected customer credit calculation inconsistencies in change orders with improved discount payload validation.
- QuickBooks Trial Balance - Fixed parser to properly handle flat-shaped CSV exports where all columns are populated in every row.