Document Extraction Frequently Asked Questions

3min

What happens when a document is uploaded for data extraction?

The provided document is automatically processed to determine its classification type and associated property(ies). For document types determined to be optimal for data extraction, SOV blanks will be populated with information.

What if I do not want the document to be used for data extraction and I just want to attach it for a submission?

Provided documents will not be used for data extraction if the system cannot link them to a property or classify them as the types determined to commonly provide impactful data for extraction. Data extraction will be automatic if documents are both linked and classified as certain types.

Does the data from the document only populate the SOV when data is missing?

Yes!  Extracted data will not override existing SOV data. Furthermore, if extracted data validates existing SOV data, then provenance will be added to the SOV.

Does the extracted data overwrite information automatically?

No.  If extracted data conflicts with existing SOV data, Archipelago structural engineers will review the conflict.

What happens if two documents have conflicting data? 

Documents are prioritized for data extraction based on their content reliability as defined by our experts. This prioritization includes document ages. Data will be extracted from documents with higher reliability. If two documents of the same reliability have conflicting data, Archipelago structural engineers will review the conflict.