PDF Remediation: How to Fix an Inaccessible Document

What Is PDF Remediation and Why Should You Care?

Picture this: you receive an important contract as a PDF. You open it on your computer, read it without any issue, and sign it. Now imagine you are a blind person using JAWS or NVDA as your screen reader. You open that same PDF and what you hear is… incoherent gibberish. Scrambled text fragments, undescribed images, tables that sound like nonsense. That is an inaccessible PDF.

PDF remediation is the process of fixing a PDF document so that it is usable by everyone, including people who rely on assistive technologies. It is not just about slapping a few tags on: it is a structural intervention that transforms a visually appealing document into one that is semantically correct.

Personally, I believe PDF remediation is one of the most neglected areas in digital accessibility. Organizations invest in making their websites accessible but keep publishing hundreds of PDFs that are completely opaque to screen readers. According to a study by Monash University (2024), over 90% of PDF documents published on government websites fail to meet the PDF/UA standard.

What Makes a PDF Accessible?

An accessible PDF is not just a file that “can be read.” It is a document that meets a precise set of technical requirements:

• Structural tagging: every element (paragraphs, headings, lists, tables, images) is marked with semantic tags, similar to HTML.
• Alternative text: all informative images have descriptive alt text.
• Correct reading order: content is read in the logical order, not the visual placement order on the page.
• Document language: the primary language is declared, and language changes are marked.
• Table of contents and bookmarks: long documents include navigable bookmarks.
• Compatible security: document permissions do not block assistive technology access.

PDF/UA: The Standard You Need to Know

Have you heard of WCAG for the web? Well, PDF/UA (Universal Accessibility) is its equivalent for PDF documents. Published as ISO 14289, it establishes the technical requirements a PDF must meet to be considered accessible.

PDF/UA-1 vs PDF/UA-2: Which One Applies?

Currently, two versions of the standard coexist:

My recommendation: if you are just getting started with PDF remediation, focus your efforts on PDF/UA-1. It is the most widely supported standard and the one required by most current regulations, including the European Web Accessibility Directive (EN 301 549). PDF/UA-2 is the future, but the tooling ecosystem is still maturing.

The Step-by-Step Remediation Process

Have you ever taken apart a piece of IKEA furniture to reassemble it correctly? PDF remediation is quite similar. You have a document that looks fine visually, but internally its structure is messy or simply nonexistent.

Step 1: Initial Assessment with PAC 2024

Before touching anything, you need to know what problems your PDF has. PAC 2024 (PDF Accessibility Checker) from the Swiss PDF/UA Foundation is the go-to tool for verifying PDF/UA conformance. It is free, runs on Windows, and gives you a detailed error report.

Run PAC 2024 on your document and review the report. Errors are classified into: document structure, alternative text, reading order, language, tables, lists, and metadata. Each error references the specific PDF/UA-1 clause that is violated.

Step 2: Tagging the Document in Adobe Acrobat Pro

Adobe Acrobat Pro is the industry-standard tool for PDF remediation. Its Tags panel allows you to add, modify, and reorganize the semantic structure of a document.

• Open the document in Adobe Acrobat Pro and go to Tools > Accessibility.
• Use “Add Tags to Document” (autotag) as your starting point. Acrobat will attempt to detect the structure automatically.
• Manually review and correct the generated tags. Autotagging is never perfect: it confuses headings with bold text, misidentifies lists, and fails with complex layouts.
• Verify that each heading has the correct level (H1, H2, H3…) and that the hierarchy is logical.
• Mark all decorative images as “artifacts” and add alt text to informative ones.

Step 3: Reading Order

Reading order is critical. A screen reader like JAWS or NVDA reads content in the order established by the tag tree, not in the visual order. If your PDF has columns, sidebars, or floating elements, the automatic reading order is very likely incorrect.

In Adobe Acrobat Pro, use the Reading Order panel to reorganize content blocks. Drag elements in the Tags panel until the order reflects the logical reading sequence.

Step 4: Accessible Tables

Tables are among the most problematic elements in PDF remediation. For a table to be accessible:

• Each header cell must be tagged as <TH> (Table Header).
• Data cells must be tagged as <TD> (Table Data).
• The Scope attribute must indicate whether a header applies to a row or column.
• Complex tables (with merged cells) need Headers and IDs attributes.

Tools like CommonLook PDF GlobalAccess greatly simplify this process. CommonLook offers a visual interface for mapping headers and data cells that is far more intuitive than Acrobat’s native Tags panel.

Step 5: Final Verification

Once remediation is complete:

• Run PAC 2024 again and verify there are no errors.
• Test the document with NVDA or JAWS. Navigate headings, read tables, check alt text on images.
• Use axe PDF by Deque Systems for complementary automated verification.
• Verify metadata: document title, language, bookmarks.

PDF Remediation Tools: Comparison

Most Common PDF Remediation Mistakes

Want to know the errors I come across again and again in the PDF documents that land on my desk? Here are the most frequent ones:

• Scanned PDFs without OCR: a PDF that is simply a picture of a document. Without a text layer, no screen reader can read it. Solution: apply OCR with Adobe Acrobat Pro or ABBYY FineReader before tagging.
• Fake headings: text that visually looks like a heading (large, bold) but is tagged as <P>. Screen readers cannot navigate by headings if they are not properly tagged.
• Unstructured tables: tables built with tabs or spaces instead of real cells. These cannot be remediated — they must be recreated.
• Images of text: text rendered as an image. This violates WCAG 1.4.5 and is impossible for assistive technologies to read.
• Chaotic reading order: especially in documents with multiple columns, text boxes, and overlapping graphical elements.