Back in the day when I first began working in localization, we didn’t have a translation management system (TMS) – they didn’t exist – and our LSP was refusing to use translation memory because it created “too much overhead” for the first venture into producing a customer care website in more than one language. Knowing that the advent of the internet was likely to produce masses of new content in – hopefully – all the languages in which the company operated, we took a huge risk. We found a startup with a prototype of a TMS and enough enthusiasm to make sure it would work. Luckily, it did, otherwise, things may have turned out very differently.

Today, on the cusp of the post-localization era, organizations are facing a similar challenge – or rather, a similar opportunity – to completely reinvent how they manage language. A transition from localization at the late end of the production cycle to fully embedded LangOps beckons those willing to make the move. It requires using a modern TMS – not as a simple project mailing system – but as the connective tissue for organization-wide language enablement.

Lights-out project management has been possible for many years. There are projects that have no need for any manual intervention other than for steps where language experts, such as translators or interpreters, must be involved. Yet few organizations have truly embraced process automation. Enterprises may expect no-touch solutions from their language services providers, yet our data shows a huge gap between the potential and the reality. Often, the content that companies put into the translation workflow doesn’t yet lend itself to automation. But now, with the ability to embed artificial intelligence (AI) in all its forms – machine learning, generative AI, and machine translation – within all aspects of content creation, management, and localization, lights-out should be the default.

Enterprises and startups willing to take the risk can shake off the traditional multivendor model to instead engage with a single source for all things language, with technology and artificial intelligence being an integral part of a human-at-the-core process,  whether it’s for lights-out, fully automated content transformation, or for human-only translations. The argument for sharing a company’s language service contracts across multiple LSPs has long been to mitigate the risk of “putting all our eggs in one basket.” Now one or two of those baskets are much more capable of keeping all the eggs safe – and even of nurturing them until they hatch. However, this option is a distant possibility for more risk-averse localization managers and their executives. It is akin to what my team in the late 1990s did with that first TMS – but with significantly more at stake.

As more core enterprise systems enable automatic translation as part of their platform – using generic engines without corporate terminology or style – localization leaders who may have already succeeded in centralizing and streamlining language processes now face the challenge of new types of uncontrolled and potentially off-brand content. Being able to communicate the why, when, and how of consistent language – and where it might not matter – is once again vital. 

What would you do? What WILL you do? Changes are coming, and you have many options to explore. Me? I’d assess the inputs and outputs, evaluate transformation requirements, study all the options, examine the risks, and be ready for a new adventure. In a few years’ time, we will look back and say, “Did we really manage language like that then?”