Remediate PDF: The Six Steps in Order

Updated · 4 min read

On this page
  1. What remediation actually is
  2. The six steps, in order
  3. Tagged is not the same as correctly tagged
  4. What can be automated and what cannot

To remediate a PDF is to give it the structure it was never built with: tags for headings, paragraphs, lists and tables, a logical reading order, alternative text for meaningful images, and the metadata assistive technology relies on. It is a sequence, and doing the steps out of order means redoing them. This guide walks the six steps, explains the difference between tagged and correctly tagged, and says plainly which parts a machine can do and which need a person.

What remediation actually is

A PDF is, at heart, a set of instructions for drawing things on pages. It does not have to say “this is a heading” or “read this column before that one”. Screen readers work around that gap, but only if someone writes the structure in.

Remediation is that writing. It is not a single button, even when a tool offers an automatic tag pass, because the tool can guess structure but it cannot know meaning. The six steps below are the order that minimises rework.

The six steps, in order

1. Identify what you have. Open the file and check whether it has a text layer at all. Select some text; if nothing selects, it is a scan and needs OCR before anything else. Then check whether it is already tagged, which tells you how much is left to do.

2. Tag the structure. Apply tags that describe each element: headings at the right level, paragraphs, list items, tables, figures. An automatic pass is a reasonable starting point for a text document, and a poor one for a complex layout.

3. Fix the reading order. The tag tree determines what a screen reader reads and in what order. Multi-column layouts, sidebars, captions and footnotes are where this goes wrong, because the tool follows the page geometry rather than the intent.

4. Add alternative text. Every meaningful image needs a description, and every decorative image should be marked as an artefact so it is skipped. No automatic tool can do this, because deciding what an image conveys is a judgement about the content.

5. Set the metadata. The document needs a language, a title, and the flag that tells a reader to announce the title rather than the filename. These are quick to set and easy to forget.

6. Verify with a validator. Check the result with a real validator, not the tool that produced it. PAC 2024 and veraPDF both produce a report you can keep.

A worked example: a university publishes a course catalogue. The file is a two-column layout with a table of module codes. Step 1 finds a text layer, so no OCR. Step 2 tags it automatically. Step 3 reveals the reading order runs left to right across both columns, so a reader hears the first line of column one then the first line of column two. The coordinator corrects the tag tree so each column reads in full before the next. Step 4 adds alt text to the department logos and marks the page borders as artefacts. Step 5 sets the language and title. Step 6 runs a validator, which now passes. The automatic pass did half the work; the human judgement did the half that mattered.

Tagged is not the same as correctly tagged

This is the sentence to remember. A file can carry a structure tree and still fail a screen reader, because the tree can be wrong. Headings tagged as paragraphs, a list tagged as body text, a table with no headers, a reading order that jumps around the page: all of these are tagged and all of them are broken.

That is why step 6 exists, and why the report from an automated tool matters more than the file it returns. A good tool tells you what it inferred and what it could not fix, so you know where to look. For what the tags mean in detail, see tag PDF for accessibility, and for the criteria a validator checks, PDF/UA.

What can be automated and what cannot

TaskAutomatable?Why
Detecting a text layerYesA simple property of the file
Applying a first-pass tag structureYes, with reviewTools infer headings and paragraphs from layout
Correcting multi-column reading orderNoRequires knowing the author’s intent
Writing meaningful alt textNoRequires understanding the image
Setting language and titlePartlyA tool can detect language; a human confirms
Fixing complex table headersNoRequires understanding the data
Validating the resultYesDeterministic checks against the standard

The honest summary is that automation handles the routine majority and hands you the hard minority. That split is the whole reason a triage step pays off before you start.

For the manual route in a desktop tool, the Acrobat accessibility checker explains how to read a report. If you would rather start with the automatic pass and see what it flags, TagRight tags a PDF in the browser and returns a report of what still needs a human; it does not write alt text and cannot tag a scan with no text layer, and it is a remediation aid, not a certification. The plans and pricing cover unlimited documents, and you can start with a free document first. For the shortest route to a finished document, pair this sequence with the accessible PDF checklist so nothing gets missed.

Frequently asked questions

What does it mean to remediate a PDF?

Remediation is the work of adding the structure a PDF was never given: tags for headings and paragraphs, a logical reading order, alternative text for images, table headers, and metadata such as language and title. The goal is a file a screen reader can navigate.

How long does PDF remediation take?

It depends entirely on the file. A clean text document can be tagged and verified in minutes. A dense multi-column report with complex tables can take hours, because reading order and table structure need human judgement.

Can I remediate a scanned PDF?

Only after OCR. A scan with no text layer has no text to tag, so it must go through optical character recognition first, and the OCR output usually needs correction before the tags mean anything.

Do I need Acrobat Pro to remediate a PDF?

Not necessarily. Acrobat Pro is the reference tool and gives you the tags panel and reading order tool. Browser-based tools can tag a PDF you did not author, and free validators such as PAC 2024 and veraPDF handle the checking step.

Is a tagged PDF the same as an accessible one?

No. Tagged means a structure tree exists. Accessible means the structure is correct: the right heading levels, a reading order that matches the page, meaningful alt text and accurate table headers. A file can be tagged and still fail.

Tag a PDF free