Meet-up (en): Frankfurt Book Fair 2025 - quality assurance
A summary of the discussion at the 'quality assurance table' during the meet-up held at the Frankfurt Book Fair 2025 about quality assurance and accessible publishing. Addressing the key findings and struggles.
15 October 2025
Key topics that are addressed
The main topic of the table were quality assurance tools and workflows. The moderator presented the participants a question about procedures and workflows they may have in place within the company.
Some participants shared their current and planned workflows:
- Typesetting and e-book production companies do both pre- and post-export quality assurance and even done user testing with users with disabilities for complex titles.
- Some publishers shared that they had a light quality assurance process that relied heavily on automated tools such as Ace by DAISY, even though they realized this workflow will not reveal all accessibility issues.
- Some expressed plans to shift the quality assurance focus more towards pre-export by creating structured manuscript templates and automated or AI-assisted workflows for e.g. image description and indexes.
After that, the discussion shifted towards the use of AI in e-book production on a more general level. Some participants had hopes that AI would solve many of the current accessibility concerns in a few years time. Others pointed out that in most of the national implementations of the EAA, publishers are explicitly responsible for the accessibility of the files they produce, and AI should not solely be relied upon.
Key findings, struggles
Quality assurance is understood as an important part of publishing born-accessible e-books. However, there were concerns and questions about:
- roles and responsibilities within the company;
- conflicts of interest between artistic creations and accessibility requirements, e.g. conforming to WCAG criteria for colour contrast vs. letting designers use the colour combinations they want to;
- how to get authors on board;
- the use of AI: who, when, how and which Large Language Models;
- uncertainties about assistive technologies: how to make sure the content works with different combinations of devices and software; is testing with assistive technology a good practice?
- the quality assurance for PDF files (still important especially in the academic publishing), which standards and tools to use with PDFs;
- the interpretation of WCAG criteria, e.g. the required level of language tagging;
- the national implementations and possible consequences for not complying with the requirements in different countries; how to invoke exceptions; must all supervisors in all 27 EU member countries be informed.
There was a lot of interest towards the use of AI in not only quality assurance but also in production and end use.
- How should publishers use Large Language Models and other AI applications such as machine learning?
- How can publishers regulate? How do their partners and suppliers use AI or protect the copyrights?
- Will AI eventually change the landscape of browsers, reading solutions and assistive technologies so that the end-user will be able to use even inaccessible files?