Online Event: AI Along the Content Chain: From Creation to Localization

How can organizations make content easier for AI systems to find and reuse while ensuring its accuracy and consistency across languages? This was the central question of “AI Along the Content Chain: From Creation to Localization,” an online event jointly organized by tekom Europe and the Globalization and Localization Association (GALA). The two sessions explored how structured knowledge and agentic AI are changing technical communication and localization, and why human expertise remains essential. 

One Knowledge Layer, Two Directions

In the first session, Harald Stadlbauer and Viveta Gene demonstrated how a shared knowledge layer can serve two purposes: improving AI visibility and making multilingual content more reliable.

Harald explored how terminology, taxonomies, ontologies, and standards such as iiRDS can help machines interpret technical information. Mapping relevant metadata to schemas such as Schema.org can further improve how search engines and AI-powered answer systems find and use content.

Viveta focused on translation and localization, where domain knowledge and terminology decisions often remain scattered across style guides, termbases, and individual experience. Knowledge graphs can connect concepts, language-specific labels, and regulatory constraints, making this knowledge accessible and helping teams identify inconsistencies before delivery.

The key takeaway: AI visibility and reliable localization are not separate challenges. Both can benefit from the same structured knowledge layer. As AI systems increasingly reuse content across languages, organizations may not control how their information is presented, but they can improve the quality of the knowledge those systems draw on.

Agentic AI for Technical Documentation: What We Learned Building It for Real

In the second session, Michael Iantosca presented a practical example of agentic AI in technical documentation, explaining how his team developed an automated workflow designed to integrate with the software development lifecycle and support publication within a 24-hour cycle.

The workflow brings together information from sources such as product plans, design documents, and video demonstrations. AI agents analyze this material, assess whether sufficient information is available, and generate documentation drafts. Michael reported that the process reduced the time needed to produce a new set of documentation topics from two weeks to just 20 minutes.

Automation, however, is only part of the solution. Product subject matter experts review and certify generated content, while technical editors optimize it and perform quality checks before publication. Michael also described the challenges of updating existing documentation, a task that proved more complex than expected and for which knowledge graphs were essential.

A major lesson from the project was the importance of bringing experts together to define the knowledge, processes, content models, and quality criteria that an automated workflow needs. Technical communication professionals remain central to this approach: they shape the content models and templates, govern terminology and knowledge structures, maintain quality standards, review outputs, and continuously improve the workflow.

What Does This Mean for Technical Communication?

Together, the sessions highlighted a common principle: reliable AI depends on well-structured knowledge, clear processes, and human expertise. Whether the goal is to make information discoverable to AI systems or to automate documentation production, the quality of the underlying content and its context matters.

For technical communication professionals, this is an opportunity to take a more active role in shaping how AI is implemented. Rather than focusing only on generating documents faster, they can help build and maintain the knowledge foundations that support documentation, translation, localization, and AI-generated answers.

Organizations can start small by building on existing terminology and metadata, improving the structure of their source content, and involving writers, subject matter experts, information architects, and linguists in designing and evaluating AI workflows. Automation works best when the knowledge and quality requirements are defined before the process begins.

Watch the Recording and Explore the Presentations

Members of tekom Europe can revisit both sessions and the panel discussion in the tekom webinar library. The event recording and presentation slides are available after logging in to “my tekom”. Explore the presentations to learn more about the approaches, tools, and practical experiences shared by the speakers.