How to Make PDFs Accessible: A Complete Guide
PDF documents are everywhere — government agencies, universities, businesses, and nonprofits publish millions of them every year. But the vast majority of those PDFs are inaccessible to people who use screen readers, braille displays, and other assistive technologies. If your organization publishes PDFs, making them accessible is not just good practice — under ADA Title II, Section 508, and similar regulations, it may be a legal requirement.
This guide walks you through everything you need to know about PDF accessibility: what it means, why it matters, the most common issues, and how to fix them.
What Makes a PDF Accessible?
An accessible PDF is one that can be read and navigated by assistive technologies. When a screen reader encounters a PDF, it relies on the document's internal structure to understand the content. A well-structured, accessible PDF provides all the information a sighted user would get from looking at the page — headings, paragraphs, images with descriptions, tables with headers, and a logical reading order.
The two primary standards governing PDF accessibility are WCAG 2.1 Level AA (Web Content Accessibility Guidelines) and PDF/UA (ISO 14289, Universal Accessibility). WCAG provides the overall framework for digital accessibility, while PDF/UA defines the specific technical requirements for PDF files. A document that meets both standards is considered fully accessible.
Common PDF Accessibility Issues
Most PDFs fail accessibility checks for one or more of the following reasons. Understanding these issues is the first step toward fixing them.
1. No Tags (Untagged PDFs)
Tags are the structural backbone of an accessible PDF. They tell assistive technologies what each element is — a heading, a paragraph, a list item, a table cell, or an image. An untagged PDF looks like a single undifferentiated block of text to a screen reader, making it nearly impossible to navigate. This is the single most common accessibility failure. PDFs exported from Word or other authoring tools without enabling the "tagged PDF" option will be untagged. Scanned documents converted to PDF are almost always untagged.
2. Missing Alternative Text for Images
Every meaningful image in a PDF needs alternative text (alt text) that describes what the image conveys. Charts, graphs, diagrams, photographs, logos, and icons all need descriptions. Without alt text, a screen reader will either skip the image entirely or announce a meaningless filename like "image_0042.png." Decorative images — those that add visual interest but carry no informational value — should be marked as artifacts so screen readers skip them intentionally.
3. Incorrect Reading Order
The reading order of a PDF determines the sequence in which a screen reader encounters content. In a multi-column layout, the reading order might jump from the first column to the second column mid-sentence, or read a sidebar before the main content. Headers, footers, page numbers, and floating text boxes can all disrupt the logical flow. The tag tree must be ordered so that content is read in the same sequence a sighted reader would naturally follow.
4. Tables Without Headers
Data tables need header cells (TH tags) so screen readers can associate each data cell with its column and row header. Without headers, a screen reader reads table cells in sequence without any context, making the data incomprehensible. Complex tables with merged cells, spanning headers, or nested structures require additional scope attributes to be interpreted correctly.
5. Missing Document Language
The document must declare its primary language (such as English or Spanish) so screen readers use the correct pronunciation engine. Without a language declaration, a screen reader may attempt to read English text with a French pronunciation engine, rendering the content unintelligible. If the document contains passages in multiple languages, each passage should be tagged with the appropriate language attribute.
6. No Bookmarks
Documents longer than a few pages should include bookmarks that mirror the heading structure. Bookmarks allow users to jump directly to specific sections without scrolling through the entire document. For a 50-page budget report or a 200-page comprehensive plan, bookmarks are essential for usability.
7. Insufficient Color Contrast
Text must have sufficient contrast against its background. WCAG 2.1 AA requires a minimum contrast ratio of 4.5:1 for normal text and 3:1 for large text (18pt or 14pt bold). Light gray text on a white background, or colored text on a colored background, frequently fails this requirement. Information must also not be conveyed by color alone — a chart that uses only color to distinguish data series is inaccessible to colorblind users.
Step-by-Step: How to Remediate a PDF
Whether you are working with a single document or thousands, the remediation process follows the same general steps. Here is how to bring a non-compliant PDF into full accessibility compliance.
Run an Accessibility Check
Before you fix anything, you need to understand what is broken. Use a tool like Adobe Acrobat's built-in accessibility checker, PAC (PDF Accessibility Checker), or veraPDF to scan the document and generate a report of all accessibility issues. This gives you a roadmap for remediation.
Add or Fix the Tag Structure
Open the document's tag tree and ensure every content element has the correct tag. Headings should be tagged as H1, H2, H3, etc. in a logical hierarchy. Paragraphs should be tagged as P. Lists should use L, LI, Lbl, and LBody tags. If the document is untagged, you will need to add the entire tag structure from scratch — either manually or using auto-tagging tools.
Write Alt Text for Images
Review every image in the document and write descriptive alt text. Good alt text is concise but informative — it should convey the same information the image provides visually. For charts and graphs, describe the key data and trends. For decorative images, mark them as artifacts. This is often the most time-consuming step of manual remediation, especially for documents with complex diagrams or data visualizations.
Fix the Reading Order
Review the tag tree order and rearrange tags so that content flows logically. In multi-column layouts, ensure each column is read completely before moving to the next. Move headers, footers, and page numbers to the correct position. Test by reading the tag tree top to bottom — does the content make sense in that order?
Mark Up Tables Correctly
For every data table, ensure header cells are tagged as TH and data cells as TD. Add scope attributes (row or column) to header cells so screen readers can associate data with headers. For complex tables with multiple header rows or merged cells, additional work is needed to make the relationships clear.
Set Document Metadata and Language
Set the document title in the metadata properties (not just the filename). Declare the primary language. Ensure the document is set to display the title (not the filename) in the title bar. Add bookmarks that mirror your heading structure for documents longer than a few pages.
Validate and Test
Run the accessibility checker again to confirm all issues are resolved. Validate against PDF/UA using veraPDF or PAC. Then perform a manual screen reader test — open the document with NVDA, JAWS, or VoiceOver and listen to how the content is read. Automated tools catch structural problems, but only a human can verify that alt text is meaningful and the reading order makes logical sense.
Manual vs. Automated Remediation
Manual PDF remediation using Adobe Acrobat Pro is thorough but extremely time-consuming. A skilled remediation specialist can spend 30 minutes to several hours on a single document, depending on complexity. For a government website with thousands of PDFs, manual remediation is not practical within any reasonable timeline or budget.
Automated remediation tools use AI and machine learning to handle the bulk of the work. They can add tag structures, generate alt text, fix reading order, and mark up tables in minutes rather than hours. The best automated tools combine AI processing with standards-based validation to ensure the output meets PDF/UA requirements.
The most effective approach is a hybrid: use automated tools to handle the heavy lifting, then apply human review for quality assurance on alt text, reading order logic, and edge cases that automation struggles with (such as complex multi-page tables or unusual document layouts).
Preventing Accessibility Issues at the Source
The most efficient way to have accessible PDFs is to create accessible source documents in the first place. If your team creates documents in Microsoft Word, Google Docs, or InDesign, following accessibility best practices during creation is far easier than remediating after the fact.
- Use built-in heading styles rather than manually formatting text to look like headings.
- Add alt text to images as you insert them, not after the fact.
- Use true table structures for tabular data, not spaces or tabs to align columns.
- Use bulleted and numbered list tools rather than manually typing numbers or dashes.
- Export as tagged PDF — in Word, check the "Document structure tags for accessibility" option when saving as PDF.
Even with good source practices, exported PDFs should still be validated. Authoring tools do not always generate perfect tag structures, especially for complex layouts.
Need help making your PDFs accessible?
CASO Comply uses AI to remediate PDFs at scale — tags, alt text, reading order, tables, and metadata, all validated against PDF/UA standards. Start with a free scan to see how many of your documents need remediation.