The number one challenge companies face when they go global is: How do I provide customer support in a variety of different languages. Why?
“Because there just is no good way to do it,” says Heather Morgan Shoemaker, CEO and co-founder at Language I/O. “A lot of a companies customer support content is held hostage in the CRM where their agents work to provide email and chat support and is available only in a single language.”
In the past, if a company wanted to offer multilingual customer support, they would often have to hire customer support agents who speak a variety of languages.
“That’s very expensive,” says Kaarina Kvaavik, Chief Business Officer and co-founder at Language I/O. “The average customer service agent is paid $45,000 per year and a business with 24/7 customer support needs more than just one agent per language on staff.”
If hiring multilingual customer service agents wasn’t an option. Many companies were asking their customer support agents to use Google Translate.
“But customer support questions are often messy,” Kvaavik explains. “User-generated content is full of misspellings, jargon and acronyms. Copying and pasting into Google translate just won’t work for those issues.”
Recognizing this was a need across industries, including hospitality, Shoemaker and Kvaavik came together to build a solution that would solve for all of these pain points. Enter Language I/O.
“We created a technology that gives companies the ability to use their single language speaking customer service agents and turn them into polyglots that can have conversations – in real time – with customers around the world in the customer’s preferred language,” Kvaavik added.
Language I/O currently works with three major CRM providers: Salesforce Service Cloud, Oracle Service Cloud and Zendesk. Within each platform, Language I/O has its own app that sits inside the CRM where the support agent works. When an agent determines a translation is needed, they can use the Language I/O app to retrieve a real-time machine translation in just a few seconds.
How is this technology any better than Google translate?
First, Language I/O is not a machine translation engine. It doesn’t actually translate requested text. Why not?
“We considered that, but we saw that there were dozens of excellent machine translation engines already available, and figured: why try to compete when you can partner,” Shoemaker says.
However, they noticed that some engines were better at translating different languages. So, instead of partnering with just one engine and using it for all languages, Language I/O identified which engine is best for which language. Then its software sends the data to whichever engine will do the best job translating based on the language requested.
But that’s not all.
Language I/O also layers company-specific and industry-specific terms over the neural machine translation engine ensuring that its customers’ industry and company specific terms, such as product names, are always correctly translated.
“For example, we have an online gambling company and online video streaming company as clients,” Shoemaker explains. “They both use the word ‘player’ in English a lot. But in Spanish a sports player is called a ‘jugador’ and a video streaming player is called a ‘tocador.’ The gambling company will always want us to use the term ‘jugador’ and the video streaming company will always want us to use the word ‘tocador.’ So, how do we make sure ‘tocador’ isn’t accidentally used in a translation for the gambling company and vice versa? We impose each client’s glossary of terms on top of whatever the translation engine returns to us to ensure the translation is accurate.”
But the real cherry on top is how seriously Language I/O takes security. It encrypts all Personally Identifiable Information (PII) immediately and throughout the platform and is GDPR compliant and ISO 27001 certified, which is not a given in the industry.
“Security is the first thing we worry about,” Shoemaker notes. “When a chat hits our platform from a customer, we automatically assume it contains personal data. So, we scan and encrypt any PII we might find. And we also ensure that none of the source content (or the translated version of it) persists either in our system or in any of the sub-processors we might be integrated with.”
While many international hotel brands can undoubtedly see the benefit of multilingual customer support, this technology does have other possible applications within the hotel space. For instance, it could be of benefit within a hotel’s mobile app where guests can chat with the hotel concierge or housekeeping staff, even if they don’t speak each other’s language. It could even be used by hotel management to communicate with staff members whose primary language is different from the manager’s.
While currently focused on text-based language support, Language I/O may soon be extending its reach. The company recently announced that it raised $5 million in A round funding. With such an infusion in capital, the company might even consider the possibility of instantaneous voice translation down the road – which would be a boon for front desk and concierge agents everywhere.