Who Is Responsible For Making Accessible Pdfs In Government?

Whenever a PDF is going to be the single source of truth for a wide audience, I start thinking seriously about calling in experts. If it's a one-off flyer with a couple of images and no form fields, I’ll try to remediate it myself. But the moment the document has complex tables, scanned pages, embedded spreadsheets, inaccessible charts, or legal/HR implications, outsourcing makes sense. Experts bring rigorous workflows for tagging, creating logical reading order, adding alternate text, fixing headings and lists, and running remediation tools against standards like 'PDF/UA' and 'WCAG'. They also do real screen reader testing rather than just relying on automated checks, which catches the subtleties that tools miss. Practically, I look at volume and frequency: hundreds of pages or recurring monthly reports are almost always worth outsourcing. I also factor in risk — public-facing materials, government procurement, or anything likely to trigger a complaint require a pro touch. If budget allows, I hire a remediation partner for an initial batch and ask them to produce detailed style guides and tagged templates so my team can handle simpler edits later. It saves time, keeps us compliant, and teaches the in-house team through example, which is a win-win in my book.

Okay, here’s how I test an accessible PDF in a way that’s actually usable — not just ticking boxes. I usually start with automated tools to catch obvious structural problems, because they’re fast and honest. I run Adobe Acrobat Pro's Full Check and the PDF Accessibility Checker (PAC 3). Those give me a baseline: missing tags, unreadable text (scanned images without OCR), missing language, or missing alt text errors. I keep a running checklist from those reports. After the auto-check, I move into hands-on testing. I open the Tags panel and the Reading Order tool to confirm headings, lists, and tables are semantically correct. I test keyboard navigation thoroughly: tab through links, form fields, and bookmarks; use Shift+Tab to check reverse order; and try Home/End and arrow keys where appropriate. Then I fire up a screen reader — NVDA on Windows, VoiceOver on macOS/iOS, or TalkBack on Android — and listen to the document read aloud. That reveals weird reading order, unlabeled form fields, or alt text that’s too terse or missing context. Finally, I mimic real use: zoom and reflow the PDF to 200–400% to ensure content remains readable, check contrast for text and images, and review interactive forms for proper labels, tooltips, and logical tab order. If it’s a scanned doc, I confirm OCR quality and check that text layers are selectable and read correctly. I also try exporting to accessible HTML or tagged text to double-check the semantic structure. When possible, I get a quick user test with someone who uses assistive tech — nothing beats actual human feedback. That last step always gives me the nuanced fixes an automated tool misses.

I'm pretty passionate about accessibility, and when it comes to public websites, PDFs are a frequent legal hotspot. In the U.S., the big names to watch are the Americans with Disabilities Act (ADA) and Section 508 of the Rehabilitation Act — public entities and many organizations must make digital content usable for people with disabilities. Courts and the Department of Justice have repeatedly treated inaccessible PDFs as falling under ADA obligations when they’re part of a public-facing service. There’s also an expectation to meet standards like WCAG (Web Content Accessibility Guidelines), commonly WCAG 2.1 AA, and the PDF/UA (ISO 14289) specification specifically for PDFs. Outside the U.S., governments often have binding rules: the EU’s Web Accessibility Directive and EN 301 549 set expectations for public sector bodies; the UK enforces similar requirements through the Public Sector Bodies (Websites and Mobile Applications) Accessibility Regulations and the Equality Act more broadly; Canada has federal and provincial laws such as the Accessible Canada Act and Ontario’s AODA; Australia’s Disability Discrimination Act has been used in accessibility cases too. The bottom line is that many jurisdictions either reference WCAG or have their own statutes that require reasonable steps to make PDFs accessible, and noncompliance can lead to complaints, lawsuits, or administrative enforcement. My go-to tip: treat WCAG and PDF/UA as your technical roadmap, document remediation efforts, and provide alternative formats while fixing the original PDFs.

My favorite trick is to build accessibility into the source file from the start. I usually create documents in Word or InDesign and use real heading styles (H1, H2, H3) instead of faking them with bold text. Styles are the backbone: they become tagged headings in the exported PDF and give screen readers a sensible outline to follow. After I’ve got styles, I add descriptive alt text to every image and check tables for proper header rows. When exporting from Word, I use Export -> Create PDF/XPS and ensure 'Document structure tags for accessibility' is checked. From InDesign I export to PDF (Interactive or Print) with tags enabled and then open the result in Adobe Acrobat Pro. In Acrobat I run the 'Accessibility' tool: Add Tags to Document if missing, use the Reading Order tool to fix mis-tagged elements, set the document language, and run the Full Check. For scanned pages I run OCR (Recognize Text) first, then tag. Finally I test with NVDA or VoiceOver, and I’ll tweak alt text, tab order, and headings based on what the screen reader actually says. It sounds like a lot at first, but once you adopt the same flow every time it becomes second nature.

Honestly, making accessible PDFs with images is mostly about planning and thinking like someone who navigates by sound or keyboard rather than sight. I start by treating every image as a piece of content that needs context: is it decorative, informative, or carrying meaningful text? For decorative ones I mark them so they’re skipped by screen readers; for informative ones I write concise alt text that explains what matters. If an image has lots of information (a chart, diagram, or a screenshot with labels), I add a longer description either inline near the image or via a link to a separate text description. Next I focus on tags and structure. I make sure the PDF is tagged, has a proper reading order, and that the figure is wrapped in a

tag with a when appropriate. If the PDF started life in Word, InDesign, or PowerPoint I export to tagged PDF and then fix any tag glitches in a PDF editor. For scanned pages I run OCR so text becomes selectable and readable by screen readers. I also set the document language, embed fonts, check contrast for any overlaid text, and ensure images that contain text have that text also present in real text form. Finally, I test. Automated checkers like PAC 3 or Acrobat’s checker catch a lot, but I also skim with NVDA or VoiceOver myself and try keyboard-only navigation. It takes a couple of passes to get right, but once I have a checklist I reuse it and the PDFs become much friendlier for everyone.

I get excited talking about this stuff because accessibility matters and it’s surprisingly doable with the right tools and a little patience. Start inside Word: use the built-in Accessibility Checker and actually follow its fixes — apply real heading styles instead of bolding, add alt text to images, mark table headers, set the document language, and use real lists. When you go to export, choose the PDF option that preserves document structure tags (Word’s Save As PDF can embed those tags). That step alone avoids a ton of headaches later. After that I open the PDF in Adobe Acrobat Pro for a cleanup pass. Acrobat’s Accessibility tools let you run the Full Check, use the Make Accessible Action Wizard, inspect and fix the tag tree, set reading order, and create proper form labels and bookmarks. I always test with a screen reader like NVDA (free) or VoiceOver to make sure it reads naturally, and then validate with PDF Accessibility Checker (PAC 3) to check against PDF/UA standards. If I need automated remediation, CommonLook or Equidox are solid commercial options, and Foxit or PDFTron can help in workflows where Acrobat isn’t available. Little tip: keeping a checklist for headings, alt text, language, table headers, and bookmarked navigation saves time — I swear by that when converting long reports.

I've grown kind of obsessive about making PDFs that actually work for everyone, and Acrobat Pro is the main toolkit I reach for when I want a document to be usable, not just pretty. First, there's the Accessibility tools panel — the 'Make Accessible' Action Wizard walks me through the basics: it runs OCR on scanned pages, creates tags, sets the document language, and prompts me to add alternate text for images. That step alone saves so much time when I'm starting from a scan. After that I always run the Full Check from the Accessibility Checker. It spits out errors, warnings, and manual checks so I can prioritize fixes. I use the Reading Order (TouchUp Reading Order) tool to set logical structure for headings, paragraphs, lists, and tables, and then open the Tags and Order panes to tidy up the hierarchy. For forms, Acrobat lets me name fields and set tab order so screen reader users can navigate them naturally. Little things like setting document title and language, marking decorative images as artifacts, and using the Preflight PDF/UA checks round out the work. It’s a lot of small, concrete options, but together they make the PDF genuinely accessible and testable with screen readers or validators, which is super satisfying.

Okay, here’s my go-to, no-nonsense checklist that actually speeds the whole accessible-PDF-for-ebook process — written like I’m talking to a friend over coffee. First, fix the source: use real styles in Word or paragraph/character styles in InDesign. Proper heading levels, lists, and table markup in the source mean the exported PDF comes out mostly tagged correctly. That alone shaves off hours. Export with “Create Tagged PDF” enabled, and embed fonts. Next, run a focused pass in Acrobat Pro: use the 'Make Accessible' wizard but don’t blindly accept everything — manually inspect the Tags panel, Reading Order, and the Order panel. Add alt text to images (short + long as needed), set the document language, and add a title/author in Document Properties. Proper bookmarks from headings are huge for navigation, so generate or clean them up. Final speed hacks: build a template with styles and export settings, keep a snippet library of standard alt-text phrases, batch-process fonts/optimize with a Preflight profile, and validate with PAC 3 or Acrobat Accessibility Checker. I always do a quick NVDA pass — if it flows for the screen reader, I call it done. It feels satisfying when a file that started as a messy draft works cleanly on a Kindle and for a screen reader.

I get oddly excited about OCR — it’s like giving a printed book a second life. When I work with scanned books, OCR is the crucial first step: it converts the picture of text into actual text that screen readers can read, search engines can index, and users can highlight or copy. Good OCR paired with careful layout analysis lets you create tagged PDFs that preserve headings, lists, reading order, and alternative text for images, which all matter for real accessibility. Practically, the pipeline I trust starts with cleaning the scans (deskewing, despeckle, contrast adjustments), running a strong OCR engine (commercial or open-source), and then manually fixing errors that matter most for navigation — headings, captions, and tables. For older, faded, or multilingual books, newer OCR models trained on diverse scripts make a huge difference, though handwriting and complex formulas still trip them up. Exporting as a properly tagged PDF or converting to EPUB with semantic tags gets you far toward compliance with standards like PDF/UA or WCAG. It's not magic: OCR reduces barriers dramatically but often needs human-in-the-loop for quality. I like combining automated OCR with spot-checking by volunteers or students; that mix keeps costs down while raising accessibility to a level that genuinely helps people who rely on assistive tech.

I get really into the little practical differences between file formats, so the PDF vs EPUB question is one I talk about a lot with friends when we're comparing e-books or lecture notes. EPUB is basically a zipped collection of HTML, CSS and images, which means it’s fluid: text reflows, fonts can be resized, and headings can be used for quick navigation. That makes EPUBs much friendlier for people who need larger text, different contrast settings, or rely on screen readers. EPUB 3 also supports semantic markup, Media Overlays (text-audio sync), MathML for equations, and is designed around accessible metadata and reading order — when an EPUB is built correctly, it almost feels like a web page that adapts to whatever device or assistive tech you use. PDFs are a different beast: they’re meant to preserve exact layout. That’s great when you want the page to look the same as the print version — think formatted reports, brochures, or complex diagrams. But that same fixed layout can be a barrier. If a PDF isn’t tagged properly, screen readers can’t follow the right reading order, headings might be missing, and scanned PDFs are just images that need OCR to become usable. I usually recommend EPUB for narrative books and general reading, while PDFs are fine for print-accurate materials — but only if the PDF has proper accessibility tagging and alt text. If I’m handing something to someone who relies on assistive tech, I try to provide both formats or at least an accessible EPUB.