LegalTech: Table of Exhibit Creator

I intern-ed in a law firm, and out of interest, built a couple of tools, one of which I’m showing here, to help automate processes I’ve seen law firms manually do. Note though, the code I’ve written depends very specifically on one particular process of work that I’ve observed.

Table of Exhibits Creator

Usually, a law firm may have to go through a large document, to find all the exhibits mentioned in a given statement to create a cover page.

This program is written in Python and replicates this process. It uses Textract, the Python package, to goes through a document (the Affidavit) to find input for a separate document, it creates, the Cover Page (coverpage.docx). This document contains a table of the exhibits by pulling every sentence with the word “page” in it, which is always used, whenever an exhibit is covered. It then uses Artificial Intelligence and Natural Language Processing (here, Python’s NLTK ) to identify the noun to fill out the exhibit column. It also identifies a date in that sentence. Using this information, it creates new, numbered rows for the table.

Where the noun cannot be identified, the whole sentence will be pulled out, and plopped into the Exhibit column of the table, for a human lawyer to sort out.