Making a PDF accessible means giving the document a machine-readable structural level: tags, a correct reading order and alt text. There are three ways to do this – accessible creation at source, manual post-processing and automated correction. Which way is the right one depends primarily on whether you need to create new documents or process an existing inventory.
What makes a PDF accessible?
An accessible PDF looks the same to sighted readers as any other - the difference lies in the invisible structural level. It allows screen readers and other assistive technologies to render the document in a meaningful way. Specifically, an accessible PDF needs:
- tags, which semantically mark up headings, paragraphs, lists and tables,
- one logical reading order, even with multi-column layouts,
- Alt text for images, graphics and diagrams,
- correctly excellent Table structures with headers and cell references,
- Document metadata like title and language.
The relevant technical standard for this is PDF/UA-2 (ISO 14289-2); WCAG 2.2 and the European standard EN 301 549 also apply. The article explains what these standards require in detail Understanding PDF/UA.
The topic is legally relevant due to the German Accessibility Strengthening Act (BFSG): Since June 28, 2025, it has applied to many companies whose products and services fall within the scope of § 1 BFSG - such as banking services and electronic commerce. BITV 2.0 also applies to public bodies. Details can be found in the post BFSG and accessible PDFs.
Way 1: Create accessible at the source
The cleanest way for new documents: You already create accessibility in the source program. Microsoft Word and Adobe InDesign can export tagged PDFs if the source document is consistently structured with styles, heading levels, alt text and defined document language.
Limits of this path
- Only suitable for rebuilding. Your existing document inventory – forms, reports, brochures from the last few years – remains unaffected.
- The quality varies. The export only creates usable tags if everyone involved uses the templates in a disciplined manner. In practice, many documents are created under time pressure or by changing authors - small deviations lead to incorrect structure.
- Complex layouts remain difficult. Multi-column pages, nested tables or design-intensive publications often require rework even after export.
Accessible creation at the source is therefore a useful basis for future documents - but not an answer to the question of how existing PDFs can be made accessible. A final check of the exported PDF is still necessary in any case.
Way 2: Manual post-processing
Existing PDFs can be subsequently prepared using tools such as Adobe Acrobat Pro: experts check and correct the tag tree, set the reading order, add alt text and mark up tables – page by page.
Limits of this path
- Expertise is a prerequisite. If you want to set tags correctly, you have to master PDF/UA, the tag semantics and the peculiarities of the tools. This knowledge is not available in most specialist departments and cannot be built up on the side.
- The effort increases with each document. Each page is edited individually; Scope and layout complexity determine the processing time. This is acceptable for individual, long-lasting documents - but for hundreds it becomes a bottleneck.
- Difficult to scale and difficult to be consistent. Different processors make different decisions. Ensuring consistent quality across a large inventory requires additional testing loops.
Manual post-processing – internally or via service providers – remains an option for small inventories and special cases. It is not suitable as a standard process for documents that are constantly being created.
Way 3: Automated correction with AI
The third way starts with the finished PDF, but replaces manual work with software: Accessful analyzes the document with AI support, understands text, layout and image content and automatically adds the missing structural level - tags, reading order and alt text in over 110 languages. The layout and content of the document remain unchanged; the correction only affects the structural level.
Results are typically available in seconds. For each document you will receive an audit-ready report as proof of audits and internal documentation. The AI reliably detects typical barriers; Anything that cannot be automatically and conclusively checked is shown in the protocol.
There are three access routes for operations:
- Web-App with drag-and-drop and live status – without installation, suitable for specialist departments,
- REST API for integration into editorial and output systems, including batch processing of thousands of documents and webhook notifications,
- On-Premises for operation completely in your infrastructure, air-gapped if desired.
The page shows what the process looks like in detail – from handover to the audit report How it works.
Which path suits which situation?
The three paths are not mutually exclusive – in practice, most organizations combine them. To help you decide:
| Initial situation | Recommended way | Justification |
|---|---|---|
| New documents, few templates, trained authors | Creation at source + automated final inspection | Accessibility arises where the document is created; the control ensures export quality. |
| Small inventory (individual, long-lasting documents) | Manual post-processing or automated correction | If there are few documents, the manual effort is manageable; automation is usually faster. |
| Large inventory or constantly new documents | Automated correction via batch processing or API | Manual editing does not scale; The automation processes thousands of documents in one go and integrates into existing processes. |
Especially in document-intensive industries – for example Banks, Insurance or in Authorities – New PDFs are constantly being created from output management and editing systems. There is practically no way around an automated process here.
Checklist: 5 steps to accessible document inventory
- Record inventory. Get an overview: Which PDFs are publicly accessible or part of your services? Which are constantly being created, which are archive material?
- Check current status. Test representative documents against PDF/UA-2 - for example with the free PDF check from Accessful, which delivers results in seconds.
- Prioritize. First, the documents that fall under the BFSG or – for public bodies – under BITV 2.0 and are actively used: forms, contract documents, product information. Rarely accessed archives will follow later.
- Correct. For each document group, select the appropriate path from the table above. This is ideal for large quantities automated correction with batch processing on; the audit reports document every result.
- Anchor processes. Make sure that new documents do not become a barrier in the first place: accessible templates at the source, automated auditing and remediation before publication - for example via API directly in the publication workflow - and an orderly storage of the audit reports.
In this way, a one-time clean-up operation becomes a permanent process: the inventory gradually becomes accessible, and new documents remain so from the start. You can find an overview of packages and conditions at Pricing.