Unstructured and complex documents user guide

The Generative Extraction for unstructured and complex documents capability allows you to define the relationship between fields, that is, the data points to be extracted, and field groups, that is, a means of grouping fields under a single concept for processing.

Start by identifying all data points or fields that you want to extract from the documents. Then, organize the fields into a coherent structure using field groups. The fields defined under a field group can be predicted multiple times as different instances of that field group, known individually as an extraction.

Figure 1. Defining the extraction schema

Figure 2. Example of an extraction schema