The InclusiveDocs Solution leverages the power of machine learning, with visual and natural language processing artificial intelligence. For publishers, this means taking content to the next level by customizing it and translating it in formats everyone can understand.
Once a PDF has been imported, the InclusiveDocs platform uses machine learning to determine the best AI model to use, at which point the Artificial Intelligence runs through each page of the PDF and tags its various elements in HTML, making sure to follow WCAG 2.1 AA accessibility guidelines.
The InclusiveDocs Solution provides a manual quality assurance graphic user interface for manual remediation. We utilize this interface for corrections, but also to train the selected AI model even further on this type of document. For new documents, after the AI chooses the best starting model for the document type and layout, we clone that model and train the new model further to increase accuracy and minimize manual QA efforts going forward for this type of document.
The Graphic User Interface provides the manual reviewer with a WYSIWYG editor for the various different tagged HTML elements. It also provides the ability to manually extract from the PDF by using the Box tool to draw a box around the content you wish to extract as well as copy/paste or type out any content that was missed by the AI.
The InclusiveDocs solution will easily and accurately extract tables, but it also provides a great set of tools that enable granular control over every aspect of the table. Adding column/row headers is as easy as highlighting and using the right-click menu, you can also copy and paste any HTML table or even create a table from scratch using the interface and manage the table properties. For more information on how to use the platform, visit our knowledge base.
The InclusiveDocs Solution offers a simple user interface to enable the easy addition/modification of image alt-text. This interface loads up all of the images that are in your document whether they were automatically extracted by the AI or manually uploaded. The filter option allows you to quickly see which images currently have or don’t have alt-text.
Once the alt-text has been added to the desired images, clicking on the “Update” button will trigger the regeneration of the content including the newly added alt-text.