An accessible PDF is one that people can read with a screen reader, magnifier or keyboard, in the right order, with headings, lists, tables and images that make sense. This guide walks you through making one step by step, from Word or InDesign through to checking it in Acrobat, with examples of the problems we see most often.
First, ask whether it needs to be a PDF
PDFs are hard to make fully accessible, especially on phones. The Australian Human Rights Commission’s 2025 guidelines say PDFs should meet the PDF/UA standard, and that organisations should consider more accessible formats. If the information could be a web page, a web page is usually better. It works on any device, is easier to keep up to date and is easier for search engines to find.
PDFs still make sense for things people need to print, sign or keep, such as forms, reports and agreements. If you need one, make it properly.
Step 1: Start with an accessible source document
Most PDF problems start in the original Word or design file. Fixing them there is much faster than fixing the PDF later. In Word:
- Use heading styles. Apply Heading 1, Heading 2 and Heading 3 from the Styles menu instead of making text big and bold. Use Heading 1 once, for the document title, and don’t skip levels.
- Use real lists. Use the bullet and numbering buttons, not dashes or typed numbers.
- Set up tables properly. Use simple tables with a header row. In Table Properties, tick “Repeat as header row at the top of each page” and turn off “Allow row to break across pages”. Avoid merged cells and never use tables for layout.
- Add alt text to images. Right-click each image, choose “View Alt Text” and describe it, or tick “Mark as decorative”. Our alt text guide has examples.
- Keep images in line with text. Floating images and text boxes are often read in the wrong order or skipped.
- Write meaningful links. “Read the annual report” instead of “click here” or a long web address.
- Check contrast. Make sure text colours are readable. See how to fix low contrast text.
- Don’t use blank lines or spaces for layout. Use paragraph spacing and page breaks instead, so screen readers don’t announce “blank” over and over.
Then run Word’s built-in checker: go to Review, then Check Accessibility, and fix what it finds.
Step 2: Set the document title and language
In Word, go to File, then Info, and fill in the Title. This becomes the PDF title that screen readers announce, instead of a file name like “Final_v3_FINAL.docx”. Also check the document language is set to English (Australia) under Review, then Language.
Step 3: Export correctly
- Word on Windows: go to File, then Save As, and choose PDF. Select Options and tick “Document structure tags for accessibility” and “Create bookmarks using headings”.
- Word on Mac: go to File, then Save As, choose PDF and select “Best for electronic distribution and accessibility”.
- Don’t use “Print to PDF”. Printing to PDF throws away the tags, so the PDF has no structure for screen readers.
If you design in InDesign
- Map each paragraph style to a tag, such as H1, H2 or P, in Paragraph Style Options under Export Tagging.
- Use the Articles panel to set the reading order, and tick “Use for Reading Order in Tagged PDF”.
- Add alt text to images in Object Export Options, and mark decorative items as artifacts.
- Fill in the document title in File Info.
- Export with “Create Tagged PDF” ticked, and set the PDF to display the document title.
Step 4: Run the Acrobat accessibility check
Open the PDF in Adobe Acrobat Pro and run the full accessibility check from the Accessibility tools. Fix anything it reports, such as a missing title or language. Remember that Acrobat’s checker only checks that tags exist. It can’t tell you whether they are right, so don’t stop here.
Step 5: Check the tags and reading order
Open the Tags panel and work through the document from top to bottom. Check that:
- the title is tagged H1, and other headings are H2, H3 and so on without skipping levels
- each paragraph is one P tag, not split into a tag per line or per word
- lists use list tags (L, LI, Lbl and LBody), and a list that continues onto the next page is one list
- tables use Table, TR, TH and TD tags, with header cells marked as headers
- images have alt text, and decorative images are marked as artifacts
- page numbers, running headers and footers are marked as artifacts so they aren’t read on every page
- the order of the tags matches the order someone would read the page
Step 6: Check it against PDF/UA
Run the free PDF Accessibility Checker (PAC) from axes4. It tests the PDF against the PDF/UA standard and WCAG, and shows a preview of the document structure. It finds problems Acrobat’s checker misses, such as incorrect table structure and missing alt text on certain elements.
Step 7: Test with a screen reader
Open the PDF in Adobe Acrobat Reader and read it with a screen reader such as NVDA, which is free. Listen for:
- text that is read in the wrong order, or words that run together
- headings and lists that are announced correctly
- tables where each number is read with its column and row heading
- images that are described, and decorative ones that are skipped
- links that make sense on their own
Also tab through the document with the keyboard to check links and form fields are reached in a sensible order.
Common problems we fix in PDFs
These come up again and again in documents we remediate, including ones that passed an automated check:
- A tag for every line or word. The screen reader pauses in odd places, and searching or copying text breaks. Each paragraph, heading or table cell should be one tag.
- Words that run together. Some software positions words on the page without real spaces, so a screen reader reads “Totalrevenuefortheyear” as one word.
- Bullets read as strange symbols. Bullets from symbol fonts can be read out as “private use character” or a random letter instead of a bullet.
- Lists broken across pages. A list that continues on the next page is tagged as two separate lists, so the count is wrong.
- Several H1 headings. Every section heading is tagged H1, so the structure is flat and hard to navigate.
- Repeated table headers. A table that runs across pages repeats its header row, and each repeat is read as more data.
- “Continued” headings. Headings like “Results (continued)” are read as new sections.
Fillable PDF forms
Every form field needs a description (called a tooltip in Acrobat) that says what to enter, such as “Date of birth, DD/MM/YYYY”. Check the tab order follows the form from top to bottom, and that required fields and error messages are clear. For many forms, an online form is easier for everyone to use.
What about scanned PDFs?
A scanned PDF is just a picture of a page. A screen reader can’t read it at all. First, run text recognition (OCR) in Acrobat, then check the recognised text carefully for errors, then tag it like any other PDF. If you have the original file, it is almost always quicker to start again from that.
Quick checklist
- Is there a document title, and is the language set?
- Is there one H1, with headings in order?
- Are lists, tables and images tagged correctly?
- Are headers, footers and page numbers artifacts?
- Does it pass PAC, and does it make sense with a screen reader?
Need help with your documents?
We fix existing PDFs, Word and PowerPoint files so they work with screen readers, without changing how they look, and test every one with JAWS, NVDA and VoiceOver. Most jobs are back in 3 working days. Find out more about our PDF remediation service or contact us for a fixed quote.