Configure sampling risk techniques
Choose the risk techniques to apply to the selected population. Review the inputs for each enabled technique before starting the run.
Choose built-in techniques
| Technique | What it flags or evaluates |
|---|---|
| Back dated | Transaction dates earlier than their posting or entry dates. |
| Closing entry | Journal entries on or around period-end close. |
| Cutoff | Transaction and posting dates on opposite sides of period end; optional day and minimum-amount inputs narrow the check. |
| Days outstanding | Open items older than the configured number of days. |
| Duplicate transaction | Repeated invoice-number and vendor or party combinations. |
| Forward posting | Transaction dates later than their posting or entry dates. |
| Keyword search | The specified comma-separated keywords in transaction details. |
| New customer | Recently created customers or vendors with activity near year-end, using the configured days and minimum amount. |
| Non-working days | Entry or posting dates on a weekend or a U.S. public holiday. |
| Outlier detection | AI review of posting, account, amount, and related anomalies. |
| Round sum value | Round-number amounts at or above your chosen minimum. |
Select a technique in the list, turn it on or off, and complete its required inputs. Add a Note when you need to record context for the selected criterion.
Distinguish risk scores from unusual items
An enabled criterion adds to a transaction’s risk score when the transaction meets it. Multiple hits contribute to the combined score.
Turn on Include as unusual item to place qualifying items in the unusual group. Unusual item threshold narrows that treatment by amount. Items below that threshold can still receive a risk score when they meet the enabled criterion.
Read the on-screen information controls for the current grouping and test-selection behavior. A zero score means that no selected criterion flagged the item; it does not prove that the item is correct.
Add a custom criterion
Use + Add custom criterion to add a check. Enter Name and Description — what should the AI flag? Write a specific condition the AI can evaluate from the available data.
Enable the criterion and configure unusual-item treatment, threshold, and notes if needed. Remove deletes that custom criterion from the current setup.
Run the sampling
Review the source, period, thresholds, and enabled criteria. Keep at least one technique or custom criterion enabled, and complete the name and description for each enabled custom criterion. Click Run Sampling once the population retrieval is complete and the setup is valid. Follow the progress indicator and any error message.
Continue with results and export.