AI automation that cut court-document processing time by 80%.
The team at All Island Media were spending eight hours each week printing then separating a weekly 800+ page court document into it’s individual cases. They would then go through each case painstakingly searching and highlighting key bits of information hidden in the text, before manually entering all the data into the CMS of a very dated website.
This manual process was time-consuming and error-prone. They needed a more efficient way.

The Challenge
The PDF that the AIM team download from the Jersey courts contains all sorts of case documents (property, debt, loan), from different lawyers, in different formats, all merged together.
We needed a method to split the document down into each individual case, identify what type of case it was, extract the required detail from each case, then finally spit out all the data in a structured way to feed the website.
The Solution
We needed a variety of tools to perform each action and then an automation tool to create the workflow that linked them all together.
We implemented a comprehensive solution including:
• PDF splitting: We used a PDF tool to read and split the original court documents when it found certain text phases. (PDF.co and Google Drive)
• Case categorisation: Each individual document would be categorised by the type of case (property transaction, divorce etc.), named and stored accordingly. (ChatGPT API)
• AI data extraction: An AI bot would extract the required information, for example in the case of property transactions: house name, vendor, buyer, cost etc. (ChatGPT Assistants API)
• Data import: The information would be structured to meet needs and a feed provided to the new Jersey Confidential website. (JSON, WordPress)
![]()
The Results
Stop working harder...
Start working smarter.