How Equidox Uses AI in Every Accessibility Solution

AI (artificial intelligence) has taken the tech world by storm. WIth the recent introduction of ChatGPT, many in the accessibility field are wondering whether similar technology can be applied to PDF accessibility. Can AI make your digital content accessible? 

The answer is yes. Equidox has been implementing machine learning–a broad subset of AI–for years, and continues to develop the technology to improve remediation speed and accuracy. The goal is to approach a fully automated process.

While most PDF accessibility cannot be fully automated out of the box, Equidox has been able to automate many components of our SaaS software. We are also able to fully automate our customized solution for high-volume PDFs, called Equidox AI. Each step taken towards automation saves time, money and resources while making PDF content accessible and usable. While some aspects of PDF remediation for unique, ad hoc documents must always be manual, the integration of AI features can significantly reduce many of the manual processes required to tag PDF elements. 

Implementing AI in PDF Accessibility

David Freelan is Equidox’s lead data scientist. 

Here’s what David has to say about Equidox’s artificial intelligence development:

“Our machine learning technology aims to continually improve the PDF remediation process. Our goal is to continue to reduce the amount of time remediators spend tagging document elements.”

“Equidox software has always been effective at identifying non-visual elements like text.The introduction of AI functionality allowed Equidox to recognize and analyze visually structured and complex data, such as lists and tables. When we first introduced AI into our software, our initial focus was on identifying and remediating tables because these elements are some of the most time-consuming for remediators. Lists followed shortly after.”

Equidox AI- Full automation

“We currently have the ability to completely automate the PDF remediation process using our tailored Equidox AI, our solution for high-volume PDFs. This solution uses a trained AI model to automatically detect specific headings, text, tables, and lists in repetitive, large volume documents. It can even be trained to apply alt text to identified images. This technology is batch-specific and is tailored for each unique document template. It is effective for system-generated documents with predictable, repetitive elements. The solution is uniquely able to “learn” what style elements each document uses to differentiate sections and features. That allows it to correctly identify even outliers in which some sections don’t look like the rest. An example might be a bank statement in which the list of charges will be longer or shorter from customer to customer. Equidox AI is ideally suited for bank statements, insurance claims summaries, utility statements, billing and customer data reports.”

Visit the Equidox AI page to learn how our technology automates PDF accessibility and removes the manual work of PDF remediation.

Equidox Software- Automated features

Equidox software features a number of AI-powered, automated features which the user guides. These features take the tedious, manual effort out of PDF remediation for accurately tagged pages in just seconds. Learn more on our Equidox Software page.

Smart Table Detector

David describes the Equidox Smart Table Detector, “We successfully implemented a machine learning detection process for tables.” 

“The Equidox Smart Table Detector uses ten settings: five for columns and five for rows. Each uses a different detection functionality to place row and column delimiters [usually gridlines] in order to tag the individual cells of the table. These work independently, so are useful for a wide variety of table formats. Some detect existing lines, some detect text, some detect spacing. By choosing one setting from the options for rows, and one setting from the options for columns, the machine learning detection process can more accurately pinpoint how to tag the content into cells.” 

“Our remediators have commented that the new Table Editor features reduced the time spent on page-sized tables from 15-20 minutes down to an average of two minutes with our new machine learning technology.” 

Smart List Detector

David explains the Equidox Smart List Detector: “Equidox is also able to detect list items automatically, including lists that have a hierarchical structure, such as nested lists. The idea is that you will be able to simply highlight the list, and machine learning will give its best guess on where the list items, sublists, and delimiters are.” This AI technology reduces list tagging to a few clicks. Some lists can be tagged in seconds. 

Page Match

Equidox’s newest unique use of AI technology is Page Match. As its name suggests, Page Match suggests a tag zone layout to use on the current document page by using machine learning analysis of previously remediated, validated pages. 

AI makes automated PDF accessibility possible 

Equidox has harnessed the unique ability of AI to make PDF accessibility faster, easier, and more automated. While training AI to automate many aspects of PDF accessibility, Equidox prioritizes accuracy and usability as well as accessibility. It is not enough to just tag documents; they must be tagged correctly.

David shares a glimpse into the future, “As our machine learning technology improves, we hope it will be able to take that final step to using artificial intelligence for near-total automation.” 

Ready to meet your PDF accessibility goals quickly and efficiently with the power of AI? Contact Equidox for more information and to schedule a discovery call.

Tammy Albee

Tammy Albee | Director of Marketing | Equidox Tammy joined Equidox after four years of experience working at the National Federation of the Blind. She firmly maintains that accessibility is about reaching everyone, regardless of ability, and boosting your market share in the process. "Nobody should be barred from accessing information. It's what drives our modern society."