Text Presence

Explanation

Text Presence checks whether specific required text actually appears in an AI system's output, which matters in document-extraction use cases where mandatory elements, such as a disclosure clause, an account number, a signature block, or a regulatory statement, must be present for the output to be valid and compliant. The Text Presence metric verifies the presence (or absence) of expected strings or content in the produced text or extracted result. To implement it, define the list of required text elements for each document type, run an automated presence check over the model's output, and log for each item whether the expected text was found, retaining the result as evidence; this check sits at the output-validation stage of the extraction pipeline and can run on every document rather than only on samples, since it is inexpensive and deterministic. The threshold is set as a target per control objective with an alert on breach (proposed), so set a risk-based threshold for the acceptable presence rate given the document type and its regulatory importance; when required text is missing beyond the tolerance, the control raises an alert that routes the document for manual review or reprocessing. Its Immediate priority reflects that missing mandatory text in a BFSI document can directly cause a compliance or processing failure.