How to make an accessible PDF, step by step

A PDF can look perfect and still be unusable for someone who’s blind. Screen readers don’t see the page; they read the PDF’s hidden tag structure. If the tags are missing or wrong, the document is silent or scrambled. Here’s how to get it right.

Step 1: Start in the source document

It’s ten times easier to make an accessible PDF from an accessible source than to fix a broken PDF. In Word (the same ideas apply to PowerPoint, Google Docs and InDesign):

• Use built-in Heading 1, Heading 2, Heading 3 styles — not bold, bigger text.
• Use the real bulleted and numbered list buttons.
• Add alt text to every meaningful image; mark decorative images as decorative.
• Build simple tables with a header row (Table Design → Header Row, and Repeat as header row).
• Write descriptive link text (“Download the 2026 fee schedule”), not “click here.”
• Check color contrast and don’t rely on color alone.
• Set the document language (Review → Language).
• Run Review → Check Accessibility and fix everything it reports.

Full walkthrough: accessible Word documents.

Step 2: Export a tagged PDF

• Word for Windows: File → Save As → PDF → Options → tick Document structure tags for accessibility. Or use the Acrobat tab → Create PDF, with “Enable Accessibility and Reflow with tagged Adobe PDF” on.
• Word for Mac: File → Save As → PDF → choose “Best for electronic distribution and accessibility.”
• InDesign: Export → Adobe PDF (Interactive or Print) → tick Create Tagged PDF; set up Articles and export tags in advance.

Never “print to PDF” — printing throws away every tag.

Step 3: Set document properties

In Acrobat Pro: File → Properties.

• Description → Title: a meaningful title, not “Microsoft Word - final_v3.docx”.
• Initial View → Show: Document Title (not file name).
• Advanced → Language: e.g., English.

Step 4: Check the tags

Open View → Show/Hide → Side panels → Accessibility Tags. You should see a tree:

• <H1>, <H2>… for headings, in a logical hierarchy
• <P> for paragraphs
• <L> → <LI> → <Lbl> + <LBody> for lists
• <Table> → <TR> → <TH>/<TD> for tables
• <Figure> with alternate text for images
• <Link> for links

If the tree says “No Tags available,” the PDF is untagged: go back to Step 2, or use Accessibility → Autotag Document as a starting point (then check everything it did).

Step 5: Fix the reading order

Screen readers follow the order of the tag tree, not the visual layout. Multi-column layouts, sidebars, pull quotes, headers and footers are the usual problems.

• Drag tags in the Tags panel into the order a sighted reader would follow.
• Mark running headers, footers, page numbers and decorative lines as artifacts so they aren’t read on every page.
• Use the Reading Order tool (Accessibility → Reading Order) to retag regions that were tagged incorrectly.

Step 6: Images, tables and links

• Images: right-click each <Figure> tag → Properties → Alternate Text. See our alt text guide.
• Tables: header cells must be <TH> with Scope set to Row or Column (Table Editor → right-click → Table Cell Properties). Complex tables need header IDs.
• Links: each link should be in a <Link> tag with descriptive text.

Step 7: Forms

Fillable forms need every field to have a tooltip (that’s what screen readers announce), a logical tab order (Page Properties → Tab Order → Use Document Structure), and accessible instructions and error messages.

Step 8: Check it — twice

1. Acrobat: Accessibility → Check for Accessibility → Full Check. Fix every failure, and review the two items it can never check: Logical Reading Order and Color contrast — they always say “Needs manual check.”
2. PAC (PDF Accessibility Checker) — free from the PDF/UA Foundation. It tests PDF/UA conformance and has a Screen reader preview that shows the structure as assistive technology sees it.
3. Listen to it. Open the PDF in NVDA or JAWS and read the first few pages, a table and any form.

The accessible PDF checklist

• Text is real text (not a scanned image), or has been OCR'd and corrected
• The PDF is tagged, and every element is in the right tag type
• Headings use H1–H6 in a logical hierarchy without skipped levels
• Lists are tagged as lists; tables as tables with header cells and scope
• Reading order matches the visual order; headers, footers and page numbers are artifacts
• Every meaningful image has alternate text; decorative images are artifacts
• Links have descriptive text and are tagged as links
• Form fields have tooltips and a logical tab order
• Document title is set and displayed in the title bar
• Document language is set (and any passages in other languages are marked)
• Text contrast is at least 4.5:1; color isn't the only way information is conveyed
• Bookmarks are included for documents longer than about 20 pages
• Security settings don't block assistive technology
• Acrobat Full Check and PAC pass, and a screen reader read-through makes sense

When to get help

Long reports, complex tables, forms, scanned archives and large volumes are where DIY gets slow. Our PDF remediation team handles single documents in days and libraries of tens of thousands of pages on a schedule.