AI Account Reconciliation
Product Overview
AI Account Reconciliation is an advanced anomaly detection integration into OneStream Financial Close Account Reconciliations, powered by SensibleAI Studio. It extends traditional reconciliation review by evaluating heuristically how each reconciliation behaves over time, using historical data to establish context and identify meaningful deviations from typical behavior.
The solution combines 16 built-in Anomaly Detectors with 21 default Data Rules to surface issues such as documentation sparsity, unexpected sign changes, reconciliation balance irregularities, changes in detail item volume and aging, and shifts in processes that can indicate emerging risk.
Results are delivered directly insight the reconciliation workflow through the AI Insights button which is used to trigger the Anomaly Detection process, and an anomalies tab that surfaces the triggering Data Rule, severity, status, and the time of the most recent detection. Because alerting logic is expressed in finance-owned Data Rules, teams retain control over what qualifies as an anomaly and tune rules and thresholds over time to align with organizational expectations.
Product Value Proposition
AI Account Reconciliations helps finance teams certify with greater accuracy and confidence by focusing attention on reconciliations that have changed in meaningful ways relative to their own history. Instead of relying exclusively on static checks, it provides behavioral context through statistically grounded metrics and interpretable detector outputs, then translates those metrics into actionable alerts using explicit Data Rule logic.
This improves prioritization during close, supports consistent remediation of issues, and reduces administrative overhead through automatic resolution when underlying conditions are corrected and detection is re-run. Optional certification controls can require anomalies to be resolved or explicitly acknowledged before certification, reinforcing audit-ready governance without removing flexibility.
User Personas
Preparers: Run AI Insights during preparation to identify anomalies early, document outcomes with comments, and resolve or acknowledge items before submission.
Financial Close Administrators: Configure which Anomaly Detectors are enabled, set look-back periods and status applicability, author and tune Data Rules, and manage certification settings and JSON-based configuration management.
Approvers: Use anomaly severity, status, and supporting context to prioritize review, request remediation where needed, and ensure anomalies are addressed before certification based on organizational policy.
Auditors: Leverage the traceable anomaly record, including the triggering Data Rule, timestamps, statuses, and comments, to support transparency and consistency in close controls and review evidence.