Run Extraction Model

The run extraction model step extracts data from records based on an extraction model.

For common document types, you might be able to use an existing Marketplace model for extraction. Otherwise, you can train extraction models in ML Studio. Extraction models are connected as modules to the run extraction model step in a flow.

Rename fields to resolve conflicts

If your flow includes multiple extraction steps, it’s possible to produce duplicate field names for common data points, like name or address. Enabling Rename fields to resolve conflicts prevents errors by renaming fields with the same name.

Duplicate field names are given a prefix with two underscores and a suffix that indicates the model type and the first four characters of the model UUID. For example, a field address extracted by a text extraction model with the UUID abcd... is renamed __address_txt_abcd.

Tip

If you enable field renaming, you can still show hidden fields in Refiner or make hidden fields visible in final results. To display a hidden field in the final results, create an additional field that echoes the hidden field. For example, create a field labeled address with the formula echo(__address_txt_abcd).