Looking for Next-Generation Localization Industry Workflows (and Workforce)

The localization industry stands at a crossroads, facing unprecedented challenges and opportunities. For many of us professionals in the field, it is crucial to examine how the landscape is evolving.  I recently wrote about the concept of vector databases, their inner workings, and their application in our Deep Adaptive AI Translation, particularly in automatic post-editing. Vector databases have emerged as a powerful tool for storing and retrieving high-dimensional data. As a CEO and solution developer, I am responsible for foreseeing the future in the cards, charting the waters, and moving strategically to adapt the company to the changing circumstances. And the company needs to develop solutions and apply them in the market.

There is a Spanish-Italian singer called Miguel Bosé (the son of a bullfighter and an Italian model, whose godfather was no less than Picasso) who, back in the early 90's renounced all his previous songs and music, never sang them again -except one- and went to produce his own music. Similarly, Jochen Hummel's welcoming Erik Vogt's dream of smashing the Trados cage Jochem built almost 30 years ago with GenAI, perfectly depicted and summarized the challenge ahead for the translation / localization industry (LocWorld 50 Silicon Valley). The trouble is that a singer like Bosé may renounce his previous works and do something new. Until recently, the localization industry had become fixated on translation memories, fuzzy matches, and word prices. It took humans' first cognitive experiences with machines producing fluent language to light the first warning lights. It took me three months to digest what ChatGPT had done, the possibilities and dangers. In my 2023 post, I analyzed some of the professions that might be worst affected by AI: Justice, Content Creation, Video Game Creation, Education and Training, Language teaching, and the translation industry.

The Investment Boom and Its Aftermath

Until 2022, the localization industry was riding high on a wave of investment. The promise of global expansion and the increasing importance of multilingual content attracted significant capital. However, this boom period has given way to a more complex reality, forcing the industry to reassess its core value proposition and future direction. One of the most significant changes in recent years has been the pivot towards data-for-AI services. This has been the driver for growth for many LSPs and is partly one of the reasons for the decline in revenue in the translation industry, according to CSA. Many language service providers (LSPs) have found a lucrative niche in providing data to train AI models, particularly for machine translation. This shift has led to a decrease in the relevance of traditional translation services as clients increasingly look for faster, more cost-effective solutions. 

However, CSA points to the first decline in the translation industry in decades and a general sense of lack of direction by many companies in such a fragmented market. The two leading research firms in the localization space, CSA Research, and Nimdzi Insights offer contrasting views on the industry's future: CSA Research paints a sobering picture. They report a decline in traditional language services, citing factors such as inflation, economic instability, and the rise of AI technologies. Their message is clear: LSPs must evolve into Global Content Service Providers (GCSPs) to remain relevant. This transformation involves moving beyond translation to offer comprehensive content creation and adaptation services for global markets. Nimdzi Insights, on the other hand, presents a more optimistic outlook. They predict substantial growth for the language services industry, potentially reaching $100 billion by 2029. This growth, they argue, will be driven by AI, immersive technologies, and multimodal communication. Nimdzi sees AI as an enabler, opening new avenues for LSPs in areas like real-time translation and virtual content creation. My good colleague Diego Cresceri deals with the general angst and confusion and analyses both points of view in his newsletter.  

The Rise of AI Translation

Except in heavily regulated industries, we're witnessing a rapid adoption of AI-powered translation solutions. These include some combination of Neural Machine Translation (NMT) and Large Language Model (LLM) post-editing, often coupled with some form of RAG for a complete automatic post-editing service. And the truth is that the results are extremely promising.

Metrics for a Japanese > Chinese automatic post-editing engine at Pangeanic, the reference being a popular, generic non-US machine translation engine

This trend is reshaping client expectations and challenging the traditional role of human translators. We are still in the infancy of the new solutions, where humans play a lesser role as measurable machine translation accuracy increases. AI Translation is not just a new way to speed production, it is a whole new language delivery system where CAT tools and even translation memories have little to say. They simply become assets to gear AI's output.

The Human Element in a Machine-Driven World

AI is undoubtedly reshaping the industry, and the role of human expertise remains a point of contention. I have talked to good professional translators who are simply leaving the industry. I have seen former sales employees leave the industry and move into IP or financials. That was very rare just 2 years ago.

CSA Research suggests that human-driven translation is becoming increasingly difficult to sell as clients demand faster and cheaper solutions. Nimdzi, however, maintains that human translators will still be essential for quality control, cultural adaptation, and managing complex workflows. This divergence in perspectives raises an interesting possibility: could we see a shortage of skilled linguistic professionals in the near future? As experienced translators retire and university programs struggle to attract students, the demand for high-level linguistic expertise might unexpectedly surge.

We face a transformation imperative: Regardless of which perspective proves more accurate, one thing is clear: the localization industry is undergoing a fundamental transformation. Within the next 5 to 10 years, assuming the current pace of technological innovation and adoption continues, we may see "Localization" evolve into something more akin to "Human Linguistic Review."

Next-Generation Workflows and Workforce

As we look to the future, several key areas will shape the next generation of localization workflows and workforce:

1. AI Integration: Successful LSPs will need to seamlessly integrate AI technologies into their workflows, using them to enhance volumes and productivity, selecting and keeping the best linguists for human expertise.

2. Specialization: With routine translation tasks increasingly handled by AI, human professionals will need to specialize in areas that require deep cultural knowledge, creativity, and complex problem-solving skills.

3. Continuous Learning: The rapid pace of technological change will require a workforce committed to lifelong learning and adaptability.

4. Multidisciplinary Skills: Future localization professionals may need to combine linguistic expertise with data science, UX design, or content strategy skills.

5. Quality Assurance: As AI-driven translation becomes more prevalent, there will be an increased need for sophisticated quality assurance processes that combine human judgment with data-driven insights.

6. Cultural Consultancy: Beyond language, professionals who can provide nuanced cultural insights and adaptation strategies will be in high demand.

7. Ethical AI Management: As AI plays a larger role in translation and content creation, there will be a need for professionals who can ensure these systems are used ethically and responsibly.

Conclusion

The localization industry is at a pivotal and transitional moment. While the challenges are significant, they also present opportunities for innovation and growth. Some translation clients are choosing to adopt machine translation straight from large AI Virtual Assistant suppliers. This sends shivers down the spine of many language companies' CEOs. But it is not as easy as connecting to an API. A symbiotic relationship between human expertise and AI capabilities will likely characterize the next generation of localization workflows and workforce. For professionals and companies in the industry, the key to success will be embracing change, investing in new skills, and continually reimagining the value they bring to clients in a rapidly evolving global marketplace. Those who can navigate this transformation successfully will not just survive but may thrive in the new era of global content services.​​​​​​​​​​​​​​​​