This article was originally published in the November 2023 issue of MultiLingual magazine.

“You’ve got to be pulling my leg,” you say, but what does that actually mean? The answer to this question, though simple it may seem, is at the core of everything we do in localization. It highlights the need for transcreation in an increasingly global economy — but that’s the old model of localization.
“Ready or not, here I come!” — Generative AI

With the advent of Gen AI, the entire translation industry is on a collision course. Today’s IT and marketing leaders face a decision to either adapt to the new world of Gen AI or rigidly adhere to the old world, yielding consequences for localization teams and language service providers (LSPs) across industries.

Enter the groundbreaking partnership of Centific and VMware, two tech giants using Gen AI to revolutionize the way businesses connect with audiences around the world.

The Centific-VMware partnership

Centific (formerly Pactera Edge) and VMware are 14-year partners in pushing the boundaries of localization possibility. This partnership — rooted in a shared commitment to ongoing innovation, cost-effectiveness, and continuous quality improvement — has resulted in a nearly unmatched capability to research the potential for AI to transform the localization industry.

LocWorld 49’s Process Innovation Challenge (PIC)

In early June of 2023, Wei Zhang (Centific) and Erji Wang (VMware) teamed up to compete in the PIC at LocWorld 49. Their project, titled, “AI-Powered Localization Source Content Analysis Tool,” demonstrated how Gen AI can generate often-missing mission-critical information — like metadata, tone, writing style descriptors, and contextual terminology data. By offering practical insights into the ability of Gen AI to improve the quality and accuracy while reducing costs and accelerating processes, it earned the duo second place in the competition.

Centific and VMware push forward, embodying a shared passion for innovation by consistently integrating cutting-edge AI technologies. This strategic partnership leverages advanced machine learning (ML) algorithms and natural processing (NLP) techniques to revolutionize translation and localization.

By combining their expertise and passion, Centific and VMware drive the next generation of intelligent localization solutions and empower enterprises to authentically engage with global audiences and break down linguistic and cultural barriers.

Wei Zhang from Centific and Ergi Wang from VMWare begin speaking at 42:25.

Joining forces for a better future

Strong client collaboration is a key component of Centific’s evolution over 25 years. As the industry experiences its inevitable upheaval — already present in other industries — and the impact of large language models (LLMs) and other Gen AI tools gain traction, this collaboration grows in importance.

Recently, McKinsey estimated that Gen AI will add $2.6–4.4 trillion annually across 63 unique use cases. Even amid slowed growth and economic uncertainty surrounding the localization industry, we can’t help but wonder how much of that growth will be specific to localization or language services.

OpenAI claims that 76% of all translation tasks will be affected in some way by LLMs. Note the “in some way.” Of course, our industry will change — that’s what better tools do.

We’d contend that a similar percentage of all translation tasks were affected by the introduction of translation management system (TMS) technology in the 1980s. Despite temporary upheaval, the industry learned and adapted to the new technology and eventually came to rely on it to enhance human skills.

Cooperatively identifying both the challenge and the solution

The key challenge is to understand what things will change and how those changes will manifest. For the past 25 years, we’ve been working closely together with clients to build roadmaps for adopting new technologies with as few potholes as possible. The following table highlights the most pressing opportunities we’ve identified regarding Gen AI and the solutions we have already implemented through this collaboration:

The key takeaway from the past decades is that localization professionals need proactive adoption strategies be competitive during major technological shifts — like Gen AI. Understanding the opportunities offered by these technologies and remaining realistic about their potential outcomes are business imperatives.
But none of us can do it alone. The supply chain must change with new practices end to end for both clients and suppliers.

Quality untethered from source content

How will we measure localization quality in the future? Traditionally, it’s broadly been determined by the target’s adherence to the source content, linguistic appropriateness, and reproduction of previous translations in a TMS to ensure consistency. As the industry evolves, global content generation and publication will require the redefinition of the criteria that constitute “quality.”

Rather than comparing to source content, localization teams will need to analyze independent language analytics, sentiment, and adherence to the brief. Think of it as source validation in 40 languages. But what does this mean for quality?

We believe that rather than language quality, localization teams will need to consider language value. That is, not just the message but the impact on the end user. The value of published language is its ability to successfully convey the client’s message in an effective voice. This will become increasingly vital as the industry moves to the hyper-personalization of user experiences.

Opportunities exist to impact quality at multiple workflow phases. Providing methodologies for benchmarking each phase, establishing feedback loops for reinforcement learning, and exploiting opportunities to impact quality before publication enables flexible and dynamic environments required for rapid response to technology and customer sensibilities.

Hello, subject matter experts (SMEs)

LLMs, being language models, respond best to natural language. As such, the ability to operate and enhance the next generation of language-based tools is an easily achievable degree of upskilling for translators looking to evolve their roles to that of the linguistic SME. This is likely the new normal as AI refocuses from traditional localization to authentic international experiences.

In AI-driven localization, SMEs’ ability to optimize and improve model performance accelerates both performance and quality. As LLMs and other tools expand to deliver comparable performance in all language markets, already achieved in the initial American English markets, linguistic SMEs will need to validate, train, and operate these emerging global tools.

The future of the localization specialist’s workflow means less direct translation and transcreation and more strategic elements like cultural appropriateness, legal compliance, and general tone of voice.

We argue that localization always has been a creative skill, but industrial evolution has restricted that creativity. Translation tasks may “be impacted in some way” to increase language experts’ work satisfaction.

Rather than sentence-level work in some complex TMS, linguists will find opportunities to own client messaging at the document publication level. Creative LSPs will have freedom to meet client ambitions with reworked and fine-tuned messaging.

Future-proof your workflows now

The sun is rising on the new world. Gen AI’s transformative power for global content publication is undeniable, and tools like Centific‘s are revolutionizing the industry, redefining traditional processes, and opening doors to new possibilities.As the industry landscape evolves and LSPs deploy new practices, new solutions will be needed. By obtaining client support for the implementation of new technologies in production environments and by modifying work practices, we can maximize the opportunity for success.