Split a scanned bundle into separate documents

The other side, or the client, has sent one long PDF that is really many documents: a run of letters, a contract with its schedules, a bundle of exhibits. Cut it into the documents it contains before anyone works on it, so each event in the chronology and each row in a review table points to one document, and the AI reads each one on its own.

The question that decides where you start is where is the bundle going? Into the case timeline, you split it as you import it. Into an insight table, you split it from its row.

Open the split

Importing to a timeline

1. In the timeline, click Import in the left-hand panel. On Emails and files, add the PDFs and click Upload.

2. Under Processing Mode, choose Manual and click Upload. Manual is basic processing without AI, and it opens each uploaded PDF of more than one page for splitting once it has been uploaded.

   

In an insight table

• Right-click the row of the PDF and choose Split document. The option is there for PDF documents only.

Mark where each document starts

1. The Split document window shows the pages. Click the scissors in the gap between two pages wherever a new document starts: the cover letter of each exhibit, the first page of each email. Click a cut again to remove it, or click Clear all cuts to start over. Fit and the page arrows help in a long bundle.

   

2. Click the counter next to Clear all cuts, which shows the cuts and the documents they make, to open Result. It lists each document with its page range and page count. Click a document to jump to its first page and check that the cut is in the right place.

   

3. Click Save 3 documents; the button counts the documents your cuts make. If you uploaded several PDFs to a timeline, the window moves on to the next one (File 2 of 4). Click Skip this file to leave a PDF whole; with no cuts, the button reads Save and next (Save on the last file) and moves on without changing anything.

4. In the timeline, each document is now its own event; in the insight, its own row. The new events start with the bundle's date and title, the later parts numbered (1), (2) and so on, so give each one its own date and title, as in add the key events by hand. In the insight, the new rows are analysed like any other document.

Tips from practice

• Cut before the AI reads. Split the bundle before you run JuristIQ on it or add columns in an insight, so every answer and every event comes from one document.
• A saved split is final. The parts replace the bundle in the timeline or the table, and a file you have saved cannot be reopened in the split window. If the bundle as served matters, for a disclosure list for instance, keep a copy of the original outside the matter.
• Scanned pages need no text to be cut. You can split first and scan afterwards; see make scanned documents searchable.

When the chronology is to be built from the documents with AI rather than by hand, see build the chronology from the case documents.