Concept: AI Bank Reconciliation
AI bank reconciliation features leverage machine learning (ML) models to improve the accuracy and efficiency of bank statement reconciliation. The models make suggestions based on patterns from your historical manual reconciliations of bank statement lines and reconcilable items.
You can enable AI bank reconciliation features at the bank account level for bank accounts that use automatic reconciliation with matching rule sets.
Exception Matching
AI Bank Reconciliation: Exception Matching assists you in reconciling bank statement lines and reconcilable items that remain unmatched after automatic reconciliation runs. You can enable this feature to:
- Generate AI matches for bank statement lines and reconcilable items.
- Review AI matches and distinguish the most likely candidates based on confidence levels and provided explanations.
- Analyze why Workday surfaces recommendations by drilling into the AI matches.
- Accept suggestions individually or in bulk to reduce time and effort.
Suggestions can have 1 of these AI match statuses:
Status | Considerations |
|---|---|
Accepted
| You've accepted the suggestion. |
Not Available
| You can no longer accept the suggestion because it contains a bank statement line or reconcilable item that you've already reconciled. |
Suggested
| You haven't reconciled any of the bank statement lines or reconcilable items for this suggestion yet. |
Exception Matching Tasks and Reports
After you enable Bank Reconciliation: Exception Matching, Workday generates AI matches when you run automatic bank reconciliations. You can review and reconcile AI matches on these tasks and reports, secured to the
Process: Bank Reconciliation
domain in the Banking and Settlement functional area:Tasks and Reports | Considerations |
|---|---|
Bank Reconciliation AI Matches
| Enables you to:
|
Manual Bank Reconciliation
| Enables you to review and accept AI matches.
When you enable exception matching, Workday adds these options to the Bank Statement Lines and Reconcilable Items sections:
|
Mass Bank Reconciliation from AI Matches
| Enables you to reconcile multiple AI matches at once.
The task displays only AI matches with a status of Suggested . |
Quick Reconcile
| Enables you to review and accept a single AI match.
You can access this task from the related actions menu of reconciliation groups with a status of Suggested in the AI Matches Reconciliation Group column on the Bank Reconciliation AI Matches report. |
Rule Orchestrator
AI Bank Reconciliation: Rule Orchestrator assists you with the discovery and creation of 1-to-1 matching rules for bank reconciliation. You can add AI matching rule suggestions to your matching rule sets to improve the rate of successful automated reconciliations on subsequent automatic bank reconciliation runs.
AI-suggested matching rules can have 1 of these statuses:
- Accepted
- New
- Rejected
Rule Orchestrator Tasks and Reports
Tasks and Reports | Security | Considerations |
|---|---|---|
Bank Reconciliation AI Rules Audit Log
| Tenant Non-Configurable domain in the Report Execution Group functional area.
To view data in the report, you also need access to these domains in the Banking and Settlement and Financial Accounting functional areas:
| Enables you to audit and analyze AI-suggested matching rules of all statuses. |
Review AI Suggested Matching Rules
| These domains in the Banking and Settlement and Financial Accounting functional areas:
| Enables you to:
|
Run AI Matching Rule Suggestion
Schedule AI Matching Rule Suggestion |
| Enables you to generate AI matching rule suggestions. |
Limitations
- Workday trains ML models on your unique data only in Preview and Production tenants. In non-Production tenants, Workday doesn't use ML models, and suggested matches don't use insights from your historical data.
- It can take up to 2 weeks for Workday to start making suggestions, as Workday trains models using data that's refreshed from your tenants weekly.
- Workday makes suggestions for bank statement lines or reconcilable items created within the last 12 months.
- The volume of historical manual reconciliations impacts the quality of AI suggestions.