Unstructured Data

Unstructured data is anything that does not fit a column-and-row schema: Office documents, PDFs, images, call recordings, chat transcripts, support tickets, meeting notes, code repositories, email archives. Estimates put unstructured data at 80–90% of enterprise data volume. For privacy, unstructured data is the harder half — personal data is embedded in free-form text rather than tagged by column name, and classifying it reliably requires content inspection rather than schema inspection.

Modern DSPM and DLP tools apply ML-based classification to unstructured content: text analysis for PII patterns in documents, image classification for ID documents, speech-to-text plus classification for call recordings. The output of unstructured discovery flows into the same data inventory as structured sources, but with different confidence levels and different remediation paths. For DSR fulfilment specifically, unstructured search is where the operational cost compounds — finding every mention of a Data Principal across hundreds of thousands of documents and recordings is not a parameterised query.