Case Study: The National Archives achieves automated records classification with RecordPoint

The National Archives achieves 74.5% training accuracy with RecordPoint

The National Archives (TNA), the official archive for the UK Government, faced the challenge of managing and preserving an ever-growing volume of digital records. To appraise and select data for permanent preservation, they recognized the future need for artificial intelligence. TNA engaged RecordPoint and its Classification Intelligence service to be part of their AI for Digital Selection project, which focused on duplicate detection, entity extraction, and classification.

RecordPoint implemented its machine learning capabilities by training a model on TNA's labelled data to automatically classify documents. The solution processed both labelled and unlabeled datasets, identifying candidate records for preservation and duplicates for disposition. The project achieved a test accuracy of 71.8% and successfully extracted entities such as organizations and people. RecordPoint's tools helped TNA increase its capabilities for leveraging AI to manage its vast collection.

The National Archives (tna)