When two Flemish tech minds reunite across the airwaves, you know it’s going to be a great conversation.
In a special all-Belgian episode of The Agile Localization Podcast, host Stefan Huyghe welcomes Jourik Ciesielski, the CTO of Yamagata Europe and Founder of C-Jay International, for a deep dive into the state and future of localization.
Jourik isn’t just a tech lead; he’s a vocal advocate for reshaping the localization industry with smarter tools, stronger engineering, and a much-needed mindset shift.
Together, they tackle the big questions: What should quality really mean in a GenAI-driven world? How do we move beyond static rule-based QA? Why is it no longer enough to just translate words when you can shape entire personalized experiences?
Let’s unpack what’s coming next and why it matters.
Listen to this episode of The Agile Localization Podcast on:
GenAI Is Just Getting Started
Jourik kicks things off with a balanced take on generative AI: yes, it’s revolutionary, but no, it hasn’t quite arrived in full force within localization.
Why? Because the industry has a change aversion problem. But that’s not the end of the story. In fact, it’s just the beginning.
Jourik believes GenAI’s biggest value lies not in replacing human translation, but in expanding what’s possible across the entire localization workflow, especially upstream. Think content optimization, content classification, routing, and even generating release notes or blog posts from technical documentation.
It’s time to stop thinking of AI as “just another translator” and start seeing it as a partner in every stage of the content lifecycle.
Rethinking Quality Management from the Ground Up
If you’ve ever been frustrated by outdated QA rules throwing irrelevant errors at you, you’ll love this part. Jourik is on a mission to modernize quality management in localization.
“Our current algorithms for QA in TMSs are obsolete,” he says. They’re rule-based, context-blind, and prone to generating a flood of false positives that human linguists must painstakingly wade through.
He proposes a smarter alternative: context-aware AI models that can evaluate translations dynamically, using customized criteria provided by the user. LLMs can understand what quality means for your specific organization and even adapt their output based on that context. Think of it as moving from a rigid rulebook to an intelligent assistant that understands nuance.
Human vs. Machine? That’s the Wrong Question
Let’s make this clear: it’s not a debate.
“Humans and machines can and should coexist,” Jourik explains. “But the role of the human needs to evolve.”
Human linguists aren’t going anywhere, but their toolbox is changing. Jourik sees a huge opportunity for linguists to become prompt engineers, guiding LLMs with precise instructions and shaping the way AI interacts with content.
With their deep knowledge of tone, context, and audience, linguists are uniquely qualified to craft prompts that make AI smarter and more effective.
And when it comes to accountability? That’s where humans still hold the line. Until we resolve that legal and ethical conundrum, human oversight isn’t only necessary but also critical.
From Rule-Based to Outcome-Based
Jourik urges the industry to shift its mindset from perfectionism to performance.
Stop trying to get every dot and comma perfect. Instead, ask: how well does the localized content perform?
In other words, translation quality should be measured by outcomes, not checklists. Is the content driving engagement, conversions, or clarity? Then it’s doing its job, even if it’s not technically flawless.
Jourik gives a sharp example: A global company’s high-traffic web pages should be translated by skilled humans and polished to perfection. But for the remaining 70% of low-traffic content? A decent machine translation might be more than enough.
Embracing the Next-Gen Toolkit
Jourik encourages localization pros to think beyond the text. AR, VR, and, especially, voice tech are areas of explosive growth, particularly in dubbing and multimedia localization. Automated tools are getting more sophisticated. Some now support full video-to-video localization with realistic emotion and intonation. However, most TMSs still don’t let you start a localization project with a video file.
Jourik also makes a strong case for personalization. Today’s LLMs can tailor the same piece of content for a C-suite audience or a group of five-year-olds. That kind of control is what will truly differentiate localization in the years ahead.
Preparing for the Road Ahead
So, what can localization professionals and organizations do now to prepare?
Jourik’s advice is simple:
- Invest in Engineering: LLMs aren’t plug-and-play. They need prompting, fine-tuning, and custom integration. That means technical skills are more critical than ever. Even a basic Python course can go a long way.
- Master Prompting: Every comma counts. Prompting isn’t just a creative task; it’s an essential skill set. And human linguists are in the perfect position to own it.
- Shift the Mindset: Outcomes matter more than process. Don’t get stuck in perfection loops. Get curious, get hands-on, and experiment with what AI can really do in your workflow.
Conclusion
The future of localization is something we need to actively build. Whether it’s rethinking quality, balancing AI and human insight, or expanding into new media formats, one thing is clear: the old playbook won’t cut it anymore.
From smart QA systems to agentic workflows, the localization industry is on the brink of a revolution. The only question is: are you ready to evolve with it?
Jourik’s Background
Jourik is the CTO of Yamagata Europe and Founder of C-Jay International, bringing vast expertise in language technology consulting, localization engineering, and market research to the localization industry. He specializes in AI integration, quality management systems, and workflow optimization for language service providers. Jourik’s practical experience in developing sophisticated localization processes and evaluating market technologies makes him a trusted voice in predicting and preparing for industry disruptions.
Localize Without Coding
