In a globally connected marketplace, new technologies ensure customer transactions won’t get lost in translation.
The world is becoming increasingly connected, and companies in search of worldwide markets need to be able to communicate with customers in their native tongues. They’re depending on sophisticated new machine translation technologies to break down language barriers.
“When you first enter a market, the early adopters for any new product — whether it’s a personal care product or a tech product — tend to be internationally focused and English friendly, so you might think you’re doing well,” said Ben Sargent, content globalization strategist at the consulting firm Common Sense Advisory. “But that’s a thin layer of people.”
To reach 80 percent of the world’s total online population, businesses need to communicate in at least 12 languages, and to reach 98 percent, they need to translate across 48 languages. Overall, the process of localization — providing accurate and culturally specific translations of any company or product-related information, marketing materials and customer support data — is a $35 billion annual industry.
Many tech firms are betting on machine translation — computer systems that automatically scan and translate written or spoken words — to handle enormous volumes of daily communications. As companies seek to reach and connect with customers on a deeper level, however, it’s important to get it right.
Balancing speed and accuracy is the key to effective machine translation systems, said Philipp Koehn, chair of machine translation at the University of Edinburgh. “Whenever the language is formalized and structured, machine translation is a great option.”
To capture all the incredibly complex nuances of language, however, a system has to understand broad contexts, slang, euphemisms, figures of speech and variations in tone and sentence structure among different countries, regions, industries and even individuals.
“The variables are huge, and at some point it just becomes computationally burdensome,” said Koehn. “The challenge now is not so much the core insights, but coming up with algorithms that are fast enough to actually do all that processing.”
Machines understanding natural language is one of the core challenges in artificial intelligence, said Koehn. “You might not get machine translation perfectly right until you achieve artificial intelligence, but for now we want to get it good enough to be usable.”
Today’s consumers expect same-day replies to questions and complaints, especially given the fast-paced nature of social media. Machine translation can help companies respond quicker.
“Businesses are increasingly using platforms such as Facebook and Twitter to communicate with customers,” added Abdessamad Echihabi, vice president of research and product development at SDL, one of the world’s largest language technology firms. “Real-time translation of online conversations and comments will enable businesses to interact with billions, rather than millions, of people.”
There are several options for businesses looking to expand their global reach. Google Translate, one of the most popular free systems, converts text between more than 100 languages and can be integrated into mobile apps, Web interfaces and browsers. Like Microsoft’s Bing Translator, which Twitter uses to convert tweets between 40 languages, Google Translate uses what is called “statistical machine translation.”
This process uses raw data crunching to estimate the odds that certain words in one language correspond to certain words in another. The larger the database, the better the odds. It works well for basic translations, especially ones involving English, but accuracy drops quickly as grammar becomes more complex and the languages less common.
In January, Skype unveiled software that makes possible real-time voice-to-voice translation in seven languages, using neural networks, a digital approximation of the structure of cells in the brain. This rudimentary form of artificial intelligence can learn and improve over time, giving more accurate translations.
Neural networks are also part of Facebook’s instant-translation system announced in July, which lets its billion-plus users post in 44 other languages at the click of a button. Initially the system used neural networks to translate between English and German, but the company plans to implement them more broadly down the road.
eBay is also making a major effort to provide automated translation for its product listings, using the behavior of users to improve the process.
“If I can see a user did a query, clicked on a product, read the description and purchased it, I can make the assumption that the translation I showed this user is good because he bought the item,” said Hassan Sawaf, head of artificial intelligence at eBay. “We can teach the machines to translate better based on that feedback — the system is learning.”
The next step in breaking down language barriers would be a device that not only understands natural human speech but can also translate conversation into multiple languages on the spot — something, say, along the lines of Star Trek’s universal translator.
While this may still be a long way off, machine translation systems are improving steadily, said Koehn.
“There are already apps that can translate menus and street signs when you’re in a foreign country,” he explained. “It’s not going to be perfect, but we’ll get there.”
Editor’s Note: This is an updated version of this original story.