Customer Base

Savvy Tables Guide

Introduction

While our automated table extraction is powerful, it doesn’t always get it right; tables can be missed, or rows and columns can be split incorrectly. Savvy Tables tackles this by putting you in complete control with a powerful and intuitive toolkit. You can correct any error with lightning speed and minimal clicks. This includes the ability to instantly identify a table that the system missed, as well as expertly add, remove, merge, or resize rows and columns to ensure your data perfectly matches the source document.

Savvy Tables is two features in one

1

Table Editor

Edit columns and tables that have been misclassified by the models on initial processing.

2

Table Mark-Up

Mark up tables that have been missed by the models.

What is the Table Editor?

If the models have incorrectly split rows or columns—either by creating too many or not identifying enough—the editor provides simple tools to fix it. With intuitive actions, you can easily merge or delete extra rows, split a single row into multiple, and add or remove columns to match the document. The editor also allows you to simply drag to resize rows and columns, which is useful when content gets cut off. This entire experience is designed to give you precise control with the fewest clicks possible.

 

What is the Table Marker?

When our system fails to detect a table within a document, the Table Marker empowers you to identify it manually. By simply highlighting the area where the table appears, you can send it for reprocessing. This action automatically builds the table structure by adding column labels and splitting the rows for you. The Table Marker is designed for documents of any length; if a request spans multiple pages, you can apply the Table Marker to just the current page or across the entire document, ensuring you can easily capture all the necessary information.

What’s on the horizon?

There will be more improvements coming to Savvy Tables shortly…

  • Enabling Savvy Tables in free-text emails. Right now, it is limited to PDFs and Word documents.

Stay tuned!