Field-level rules
Required fields, formats, allowed values, ranges, sign conventions and referential checks.
Validation is where reporting risk is actually managed. Datox runs your checks automatically on every cycle, raises anything that fails or looks unusual as an exception, and gives the reviewer the evidence needed to resolve it.
Required fields, formats, allowed values, ranges, sign conventions and referential checks.
Consistency between related fields, totals against components, dates against periods, categories against instrument types.
Compare figures against control totals, accounting extracts, prior periods or a second independent source.
Structural and content validations specific to the reporting framework or output schema being produced.
Materiality thresholds and tolerated variances, so small differences do not drown out real issues.
Period-over-period movement and pattern checks that flag values which are technically valid but unusual.
When data is read from a document rather than a structured feed, the reliability of that read matters. Datox records a confidence indicator alongside extracted values and treats low confidence as a reviewable condition.
A rule fails, a reconciliation breaks or a value falls outside tolerance. The item enters the exception queue with its rule and context.
An investigation agent can compare against the source, prior periods and related records, and summarise what appears to differ.
A reviewer resolves the item: correct the source data, adjust the mapping, accept with a documented reason, or reject the run.
The decision, the reason and the reviewer are stored with the reporting period as part of the audit trail.
The specific rule set for each workflow is defined with you during deployment. Datox provides the control framework; the thresholds and policies remain yours.