# AI assistant interaction patterns

Source: https://developer.avalara.com/transaction-reconciliation/uiq8966737212286/

# AI assistant interaction patterns

Understand how the AI assistant supports reconciliation with proactive monitoring, analysis, and guided resolution patterns.

The AI assistant provides structured guidance throughout the reconciliation workflow, regardless of where it’s embedded, including ERP systems, the AvaTax portal, or partner applications.

The following interaction patterns define how the AI supports reconciliation:

-   **Proactive monitoring**: Continuously compares ERP and AvaTax transaction data and surfaces discrepancies before they affect tax filings or reporting.

-   **Root cause analysis**: Analyzes the underlying cause of discrepancies, including synchronization errors, timing differences, configuration issues, and duplicate transactions across channels.

-   **Guided resolution**: Presents structured resolution workflows for each discrepancy category and explains proposed corrective actions before implementation.

-   **Batch operations**: Supports bulk resolution actions, such as resolving all address validation errors or voiding all identified duplicate transactions within a defined scope.

-   **Confirmation gates**: Requires explicit user approval before modifying transactions and displays proposed changes and their impact for review.

-   **Audit trail logging**: Logs every AI-initiated action with the approving user, timestamp, and reason for the change to support audit and compliance requirements.