China’s once-hot sharing economy may have fallen well short of its promise to revolutionize how people use bikes, umbrellas and even toilet paper, but Shenzhen-based start-up Langogo still believes the model works well for its artificial intelligence-powered translation devices.

The latest venture of Zhang Yan, one of the co-founders of China’s bike-sharing pioneer Mobike, Langogo is planning to deploy its AI translators at airports and hotels where travelers can rent one by scanning a code with their smartphones.

Zhang Yan, chief executive of Langogo, is using Japan as a test bed for shared translation devices. (Picture: Handout)

Zhang’s test bed for the sharing model will be Japan, already the biggest market for the one-year-old start-up with 10,000 devices shipped since its launch there in January.

“Tens of millions of foreign visitors will be coming to Japan for the Olympic Games in 2020. There’s an opportunity to serve this huge group of users but you can’t expect them to buy a device for this trip,” said Zhang, who left bike sharing pioneer Mobike in 2015.

He added that Langogo has also partnered with Chinese property developer and hotel operator China Jinmao to test the shared translator business model in its home market.

The company’s first product, Genesis, is a pocket-sized gadget combining real-time translation, voice assistant, shareable Wi-fi and camera all in one device.

The Genesis pocket-sized translation device from Langogo. (Picture: Handout)

It is powered by the aggregated translation engines provided by Google, Microsoft, Tencent and Suzhou-based voice analytics firm AI Speech. With the help of self-developed AI technology, it can automatically detect the user’s own language and translate and read another language in real time.

“We’re developing the new features based on the basic technology of our partners, because there's no need for us to reinvent the wheel,” Zhang said. “We are more focused on how we apply the technology in real-life scenarios and commercialize it.”

Zhang’s interest in the field was ignited after he left Mobike to join AI Speech as general manager of the south China region. He wondered if there was a better way to commercialize AI technology and thought a hardware device may be the answer.

Started as a crowdfunding project on Indiegogo in 2018, Langogo has sold over 20,000 devices globally this year with Japan the biggest market followed by Germany and the US.

The company plans to target its home market later this month, though Zhang expects to face tough competition from similar services offered by China’s tech majors.

Chinese voice recognition specialist iFlytek and search engine giant Baidu, two of the more recognized names in AI in the country, have both launched their own translators in recent years. The two companies were also named alongside Alibaba Group and Tencent Holdings in the Chinese government’s first batch of national champions tasked with driving AI development in the country.

The development of AI – technologies which perform tasks that are characteristic of human intelligence, such as understanding language and recognizing objects and sounds – has been a priority among China’s national strategies.

Langogo and its rivals are eyeing a domestic language translation device market valued at 56 billion yuan (US$8 billion), driven by growing outbound tourism and international business exchanges, according to a 2018 report by ASKCI Consulting.

Zhang said he hopes to compete against the bigger players by covering more languages and by using its AI technology to automatically detect the language people use when travelling in multilingual regions such as Europe. The device supports 104 languages, with more than 90% accuracy for some languages including Mandarin, Cantonese and Japanese, according to the company.

Langogo wants to expand from the travel to business market with a plan to launch a Pro version that can identify voices coming from different directions, separate multiple voices speaking at the same time, and analyse the audio for transcription and translation.

A product like Pro would work better in a professional setting which has a more complicated environment than day-to-day communications, Zhang said.