Healthcare-provider organizations can manage large libraries of digital documents that need accessibility improvements. To examine how artificial intelligence handles this kind of backlog, we ran 20 real-world healthcare-provider PDF documents through EqualWeb's AI remediation engine.
The goal was to measure the files before and after automated remediation and quantify what changed.
The benchmark in one chart
The chart below plots the average EqualWeb PDF accessibility report score of each group of documents, before and after automated remediation, on the report's 0 to 100 scale. The figures are EqualWeb's own benchmark of 20 healthcare-provider PDF documents totalling 170 pages, published on 1 September 2026: Group A is the 10 documents that arrived with no accessibility tag structure, Group B the 10 that already carried some. The scores are automated checker output, not a compliance certificate, and the full methodology is set out in the sections that follow.
Defining the scores
All scores in this benchmark are EqualWeb PDF accessibility report scores on a scale of 0 to 100. The scores are produced by EqualWeb's PDF checker and remediation engine.
These scores represent the automated detection of accessibility elements. They are not a claim of WCAG or PDF/UA compliance. A higher score reflects fewer detected accessibility issues in the automated report.
An automated report score does not establish certified or legal compliance. Expert manual review is still required before making a compliance determination for a specific document.
The benchmark methodology
For this case study, we tested 20 real-world healthcare-provider PDF documents. Each document was 15 pages or fewer, and the files contained 170 pages in total.
To evaluate performance across different starting conditions, we divided the documents into two groups of 10 based on their existing accessibility structure. We then processed the documents using EqualWeb's AI PDF remediation engine.
The checker runs 80+ automated checks across 8 categories, including checks related to document structure, reading order, and tagging.
Group A: Starting without a tag structure
Group A consisted of 10 untagged source documents. These PDFs had no accessibility tag structure before processing. Without tags, assistive technologies such as screen readers may not reliably convey elements such as reading order, headings, or image descriptions.
The average report score for Group A before remediation was 0. An untagged source document reports a score of 0 before remediation because the checker has no tag structure to evaluate.
After automated remediation, the average report score for Group A increased to 96.4. Post-remediation scores ranged from 82 to 99. The group contained 104 pages, and the average processing time was 109.6 seconds per document.
Group B: Building on an existing structure
Group B consisted of 10 already-tagged source documents. These PDFs had some accessibility tag structure before processing.
The average report score for Group B before remediation was 84.9. After processing, the average increased to 96.8. Post-remediation scores ranged from 93 to 100, with 100 as the highest post-remediation report score in the group.
Group B contained 66 pages, and the average processing time was 71.3 seconds per document.
The headline finding: Closing the report-score gap
Documents that started with no accessibility tag structure in Group A finished with an average report score of 96.4. Documents that began with some tag structure in Group B finished with an average of 96.8.
The difference between the two post-remediation averages was 0.4 points. The starting difference between the group averages was 84.9 points, while the difference after processing was less than 1 point.
Across all 20 documents and 170 pages, the average post-remediation report score was 96.6. Every document improved from its starting report score.
The processing-time results were also notable. The measured averages were 109.6 seconds per document for Group A and 71.3 seconds per document for Group B. This benchmark measured automated processing time only. It did not compare total time with a manual remediation workflow.
What this means for organizations with PDF backlogs
Many organizations manage large PDF libraries that need accessibility improvements. Regulatory pressure on document accessibility is increasing in several markets, alongside long-standing requirements for accessible websites. This is part of why organizations with a PDF backlog are looking at automated remediation now.
In this 20-document healthcare-provider benchmark, files with no starting tag structure finished with an average report score close to that of files that began with some tag structure. Automated remediation can reduce the amount of structural work left for manual review.
For Group A, the average processing time was 109.6 seconds per document. That result describes the average measured in this sample, not a guaranteed processing time for every file or PDF library.
An honest look at automated remediation
Automated remediation can provide a fast first pass, but it is not a legal guarantee or a substitute for expert judgment. Report scores measure what the automated checker detects. They do not certify that a document complies with every applicable accessibility standard or legal requirement.
Complex tables, nuanced reading orders, and context-specific alternative text can require expert manual attention to provide an equitable experience for people using assistive technology. In Group A, the average report score increased from 0 to 96.4 before expert manual review.
An automated report score does not establish certified or legal compliance. Expert manual review is still required before making a compliance determination for a specific document.
Take the next step with your PDF library
If your organization has a PDF backlog, automated tools can help begin the remediation process and identify work that may still need expert review.
Explore EqualWeb's PDF Tools and AI Auto-Remediation features to see how they can support your document accessibility workflow. Contact the EqualWeb team to learn more about applying automated remediation to your PDF library.
Keep reading
- What is PDF accessibility and why is it important?
- Step-by-step guide to tagging PDFs for accessibility
- WCAG for PDF documents
Last updated . Reviewed by the EqualWeb accessibility team.
Have a PDF backlog of your own?
Check a document against 80+ automated tests, or talk to a certified specialist about remediating a full library.