datascalehr’s AI, KMod™, learns from the decisions of payroll experts in four steps.

First, a file arrives and datascalehr reads the file’s structure (headers, field names and data types). Second, KMod™ proposes mappings, for example “Field A likely corresponds to Field B”, based on relationships experts confirmed before. Third, a payroll expert confirms or rejects each proposal. Fourth, KMod™ stores the confirmed relationship.

The confirmed relationship is available immediately, with no retraining cycle, so a correction made in Munich improves the next suggestion in Vienna. KMod™ never stores file content or Personal Data.