Transifex CEO Chris Menier on How Language AI Is Now a C-Suite Priority

Transifex CEO Chris Menier on How Language AI Is Now a C-Suite Priority

Since the advent of ChatGPT, and subsequent open-source releases from OpenAI competitors, AI has been named as the source of, and solution to, problems across a range of industries — including localization.

Chris Menier, CEO of AI-powered translation management system Transifex, has a more nuanced approach. He shared his take at SlatorCon Remote March 2024.

According to Menier, the current paradigm — and its temporal and financial constraints — is holding the industry back. Clients are currently unable to localize and/or generate all the content that could contribute to more successful global strategies. 

So, what is the hold-up? Menier believes the human-in-the-loop workflow, currently a cornerstone of emerging AI-integrated localization, has caused a bottleneck, adding time and cost to the process. 

“A straight API to ChatGPT is not the answer, nor is just tossing your glossary in a prompt window. That’s not going to get you there,” Menier said. On the other hand, while AI-generated content will not take over all aspects of localization, “thinking that it’s not going to take over many of them, I think, is being a little naive,” he added.

Menier, whose experience prior to Transifex included more than two decades in tech, believes that AI-generated content comparable to human quality is “not a myth,” and there are already some options language services providers (LSPs) can offer their clients.

LSPs can use AI and localization to enable quality content creation at scale while maintaining a client’s brand. AI models can analyze and “learn” linguistic style from source content.

Menier is also a proponent of checking content quality to narrow down the number of strings that require human review — for instance, in the form of post-editing — to achieve “massive efficiencies.”

“In order to scale this, we need quality estimation that mimics human review and not only identifies those strings that have issues, but why they have issues,” Menier explained. (Transifex launched a quality index in private Beta earlier in 2024, with plans to release a public version later in 2024.)

Of course, humans will still be integral at specific points in the localization process, and in certain regulated industries. The key, Menier said, is identifying those junctures and developing a game plan for now, and for the future.

“AI is not ready to solve all of these problems today. But if we identify the areas where it’s likely to help in the near future […] this prepares you to quickly swap out technology without needing to redesign your entire process,” he explained. “Let’s get that framework in place and we can have humans where we need humans today.”

A Seat at the Table

An AI-enabled localization workflow can lay the foundation for producing and managing multilingual content at scale. It can also introduce localization as a priority to decision-makers in the C-Suite.

Menier recommends first formalizing internal localization practices, and ensuring that the team has a leader with the authority and mandate to centralize and innovate the process. 

Next, an audit should examine the frequency-of-use value to customers and prospects. This data strategy measures the utility of content, based on factors such as time spent on each page, to help automate decisions in the future. 

These hard numbers allow the localization team to present executives with a clear case for proposed budgets.

“Sometimes localization budgets are simply tack-ons to product or marketing budgets, where I believe that they should be centralized self-sustaining budgets on their own,” Menier noted. 

As executives realize localization is a key piece of “equipment” in a company’s flow, with bottlenecks slowing down time to market, he said, “the C-Suite is starting to want to participate in these buying decisions earlier.”