Pillar guide
PDF remediation, without re-authoring the archive
PDF remediation is the work of making an existing PDF usable by assistive technology: a real reading order, headings, table headers, alt text, form labels, document language and, for scans, a text layer. This guide covers what the standards require, how remediation is done, and how to plan it for a public archive.
What is PDF remediation?
PDF remediation is the process of adding the structure a PDF needs to be read by screen readers, braille displays and other assistive technology. A PDF can look complete on screen and still be unusable: the visual layout is paint, not structure. Assistive technology reads the tag tree, reading order, alternative text and table relationships behind that paint.
Remediation usually means some combination of:
- adding or repairing tags and the logical reading order;
- marking headings, lists, figures, captions and footnotes;
- associating table header cells with data cells;
- writing alt text for meaningful images and marking decorative images;
- labelling form fields, buttons and error messages;
- setting the document language and title metadata;
- running OCR on scanned pages so text exists at all;
- checking the result with a screen reader and an accessibility checker.
“Tagged” is not the same as “accessible”. A PDF exported from a word processor can carry tags that are technically present but wrong: every figure announced as “graphic”, every table flattened into a stream of numbers, every heading level skipped. Remediation is the review that turns a tag tree into something a person can navigate.
Who needs PDF remediation?
The short answer: any organization that publishes documents the public relies on and is covered by an accessibility law, regulation or contract. The long answer depends on the obligation:
| Obligation | Who it covers | Document standard in practice |
|---|---|---|
| ADA Title II | State and local governments, school districts, special district governments | WCAG 2.1 Level AA for web content, including published documents |
| Section 508 | Federal agencies and many federal contractors | WCAG 2.0 Level AA through the Revised 508 Standards |
| ADA Title III | Businesses and nonprofits open to the public | Effective communication; WCAG is the common benchmark in settlements |
| State laws | Varies by state (for example Colorado, California and New York rules) | Often WCAG 2.1 AA or a state-specific standard |
| Procurement | Buyers of software and documents | VPAT / Accessibility Conformance Report and contractual accessibility terms |
Most teams arrive here because of one of three triggers: a formal deadline, a complaint, or a procurement requirement. The ADA Title II guide covers the 2027 and 2028 dates; the Section 508 guide covers federal requirements.
WCAG 2.1 AA, PDF/UA and Section 508: how they fit
Three standards come up in almost every PDF remediation project, and they overlap:
- WCAG 2.1 Level AA is the web content standard referenced by the ADA Title II rule. Its principles — perceivable, operable, understandable, robust — apply to documents as web content. Success criteria such as 1.3.1 (info and relationships), 1.1.1 (non-text content), 1.4.3 (contrast) and 2.4.3 (focus order) are the working checklist.
- PDF/UA (ISO 14289) is the ISO standard written specifically for accessible PDF. PDF/UA-1 is the version most checkers target today; PDF/UA-2 (2024) aligns with PDF 2.0. PDF/UA gives a precise, machine-checkable definition of a conforming tag structure.
- Section 508 is the U.S. federal law. The Revised 508 Standards (36 CFR Part 1194) incorporate WCAG 2.0 AA by reference and add requirements for software, hardware and support documentation.
A practical target for a public archive is WCAG 2.1 AA behaviour, validated against PDF/UA structure. WCAG tells you what a reader must be able to do; PDF/UA tells you what the file must contain.
What a remediated PDF actually contains
Check for these before you accept a remediated file:
- A correct reading order. Columns, sidebars, footnotes and page numbers are read in the order a person would read them.
- Real headings. H1 through H6 are used to describe hierarchy, not font size alone.
- Tables that announce themselves. Header cells are associated with data cells, captions are present and merged cells do not scramble the reading order.
- Alt text with a purpose. Charts and photographs explain what matters; decorative rules and flourishes are marked as artifacts.
- Forms that work without a mouse. Every field has a label, the tab order matches the visual order, and errors are announced.
- Language and title metadata. The document language is set, and the title appears in the window and in screen-reader announcements.
- A text layer for scans. OCR text is accurate enough to read and search, and hidden OCR layers do not duplicate content.
- Links with names. “2027 budget” rather than “click here”, with the URL available to assistive technology.
Manual, automated or hybrid remediation?
Automation is good at finding problems and at repetitive structural work: OCR, tag scaffolding, metadata, alt-text suggestions, link naming and detecting missing headers. Automation is not reliable at judgement: which image is meaningful, whether a merged cell is a header, whether a chart needs a long description, or whether a page is decorative.
A hybrid workflow is the realistic answer for an archive:
- Automate the inventory. Crawl published documents, record page counts, text coverage, tag presence and last-modified dates.
- Automate the first pass. OCR scans, add structure, generate candidates for alt text and headings.
- Human-review by exception. Spend reviewer time on tables, charts, forms and ambiguous structure — not on re-keying every page.
- Verify with a screen reader. Automated checkers miss reading-order and announcement problems that a person hears immediately.
Pure “automated compliance” claims deserve scepticism. A tool can make remediation faster; it cannot decide what a document means.
A seven-step remediation workflow
- Inventory and prioritise. List documents, owners, page counts and usage. Start with high-traffic, high-impact documents: budgets, agendas, forms, notices, reports and anything required to apply for a service.
- Decide the target. Confirm whether you are meeting WCAG 2.1 AA, Section 508, PDF/UA or a state standard, and whether the original PDF must remain the source of truth.
- Fix the source where possible. If the document is still being authored, fix the source file first. Remediating the export is cheaper than remediating it again next year.
- Run OCR and structural repair. Recover text from scans, set language and title, repair tags, reading order, headings and lists.
- Review the hard parts. Tables, charts, maps, forms, footnotes and multi-column layouts get human attention.
- Verify. Test with NVDA or JAWS on Windows and VoiceOver on macOS/iOS, plus a PDF/UA checker. Record what you tested.
- Publish and maintain. Keep the accessible version linked from the same page, record the remediation date and re-check documents when they are updated.
The PDF remediation checklist turns these steps into a copy-ready list for reviewers.
PDF remediation for government, schools and healthcare
Most remediation programmes start in one of four settings:
- State and local government. Counties, cities, school districts and special districts publish budgets, agendas, permits, notices and forms. ADA Title II is the usual driver, with the 2027 and 2028 dates covered in the deadline guide.
- Federal agencies and contractors. Section 508 applies to published documents and flows into procurement. See the Section 508 guide.
- School districts and universities. Enrolment packets, board packets, policies and course materials must be usable by students and parents. District offices are covered by ADA Title II; universities also handle Section 504 and procurement reviews.
- Healthcare and regulated services. Patient notices, plan documents and public reporting fall under ADA effective-communication duties and, in some states, specific accessibility rules.
The common thread is not the sector — it is that the document is the service. When a form, notice or report is the only way to apply, appeal, pay or participate, an inaccessible PDF is a barrier, not a formatting problem.
What drives the cost and timeline
Remediation is priced by complexity, not by page count alone. The biggest cost drivers are:
- Text vs scan. A scanned page needs OCR and proofing before structure work begins.
- Tables and charts. Complex financial tables and data visualisations take far longer than body text.
- Forms. Field labels, tab order, tooltips and error handling all need testing.
- Volume and repetition. A hundred budget PDFs from the same template are cheaper per document than a hundred one-off layouts.
- Review requirements. Legal, communications and subject-matter review add calendar time even when remediation is fast.
For planning, split the archive into three buckets: documents that must be perfect now (forms, notices, high-traffic reports), documents that need a good accessible alternative (long archives), and documents that can be retired or consolidated. Retiring a stale PDF is often the cheapest remediation available.
Build, buy or use a remediation layer?
There are three common approaches:
- Remediate the files. The PDF itself becomes accessible. This is the strongest outcome and the most expensive at archive scale.
- Publish an accessible alternative. Keep the original PDF for people who need it and provide a structured HTML transcript with the same content. This is what Remediate does: the reader gets headings, tables, alt text, translation and citations, while the original file stays available.
- Do nothing and handle requests individually. This is the pattern that produces complaints. It also fails the “equally effective” test when a document is used to access a service.
Many teams combine approaches: remediate the forms and high-use documents, publish transcripts for the long tail, and set a review date for everything else. The right mix depends on your deadline, archive size and legal advice.
Frequently asked questions
What is PDF remediation?
PDF remediation is the process of making an existing PDF accessible to assistive technology by repairing its tags, reading order, headings, table headers, alt text, form labels, language metadata and, for scans, its text layer.
Is a tagged PDF automatically accessible?
No. Tags can be present but wrong. Reading order, heading levels, table relationships, alt text and form labels still need review and often repair.
Does PDF remediation change how the document looks?
Good remediation should preserve the visual design. Most changes are structural and invisible: tags, reading order, alternative text and metadata.
How long does PDF remediation take?
A simple text document can take minutes; a complex financial report or a scanned form can take hours. Tables, charts, forms and OCR are the main cost drivers, not page count alone.
Can scanned PDFs be remediated?
Yes, but they need OCR first. OCR creates a text layer that can then be structured and verified. Poor scans may need re-scanning at a higher resolution before the result is usable.
What is the difference between WCAG and PDF/UA?
WCAG describes what a person must be able to perceive, operate and understand. PDF/UA (ISO 14289) describes the technical structure an accessible PDF must contain. They are complementary; WCAG 2.1 AA plus PDF/UA structure is a common target.
Do we have to remediate every historical PDF?
Not necessarily. The ADA Title II rule has exceptions, including archived content and preexisting conventional electronic documents, but those exceptions have limits and effective-communication obligations still apply. Confirm your obligations with counsel and prioritise documents used to access services.
How do we verify remediation?
Test with a screen reader (NVDA or JAWS on Windows, VoiceOver on macOS/iOS), run a PDF/UA checker, and verify the reading order, table headers, forms and alt text manually. Keep a record of the tests and the date.
How do you make a PDF ADA compliant?
Confirm the standard that applies, then add a text layer for scans, repair tags and reading order, mark headings, lists, tables and figures, write alt text, label form fields, set the document language and title, and verify the result with a screen reader and a PDF/UA checker. Keep the original file and publish the accessible version from the same page.
Should we buy PDF remediation software or a remediation service?
Software is a better fit when your team will review and maintain documents continuously and wants the work in-house. A service is a better fit for a one-time backlog or for complex tables, forms and scans. Many organizations combine both: a remediation layer for the long tail and file-level remediation for high-use documents.
This guide is general information, not legal advice. Confirm your obligations with counsel and your accessibility office.