While earning his bachelor’s degree in electrical engineering at the Korea Advanced Institute of Science and Technology, Seyeob Kim realized that the biggest hindrance to developing artificial intelligence is the time-consuming task of collecting and annotating vast amounts of data, which he says takes up 80% of AI developers’ time. So Kim recruited three classmates—Haseung Jeon, Munhwi Jeon and Howook Shin—to launch SelectStar and fix the data problem. Its solution: Crowdsource it.
The SelectStar quartet are among the entrepreneurs in the Enterprise Technology Category of this year’s Forbes 30 under 30 Asia list that are using AI technologies to improve everything from customer service and drafting legal documents to job hunting and even the way AI itself is trained.
Launched in 2018, SelectStar develops mobile and web applications that pays users to upload their annotated photos, which SelectStar then uses to teach computers to recognize what’s in the photos, a form of AI companies then pay SelectStar for. So far, almost 90,000 people in Korea have uploaded photos, and Kim expects about 200,000 people will be on the platform by the end of the year, while SelectStar is already selling the results to LG, Samsung, internet giant Naver and 130 other customers. The startup raised $4 million last year from Korean VC firms, including the venture arm of internet giant Kakao, in a deal that valued it at almost $22 million.
Elsewhere in South Korea, home to some of the world’s most innovative technology companies, Seung Hwan Jeong, Hyungjun Mun and Junhyeong Park‘s LionRocket uses an AI-based text-to-speech technology that can mimic anyone’s voice after analyzing a 20-minute voice clip. Founded in 2019, the South Korean startup’s technology can be used to produce audiobooks and videos narrated by celebrities. In September, LionRocket formed a strategic partnership with Korean conglomerate SK Group, a leader in semiconductors and electric vehicle batteries, to commercialize its technology.
Meanwhile in Bangladesh, Mir Sakib‘s Cramstack uses AI to automate extraction of information from unstructured data to provide insights for businesses. Sakib got the idea to build a search platform that would allow users to search enterprise data sources as easily as a Google search while working at a pharmaceutical company where he struggled to find such data. Cramstack also offers tools to extract and process data from PDFs and images. It has clients in power, finance, manufacturing, healthcare and retail sectors including the government of Bangladesh, BCG, UNDP and National Bank. During the pandemic it provided government data from healthcare workers and immigration officials to help track, and curb, the spread of the coronavirus. It has raised over $1 million from Rockstart (Netherlands), Grameenphone (unit of Telenor) and angel investors.
In neighboring India, Bharath Rao‘s Precily developed an AI tool that can summarize corporate and legal documents within seconds for tax and law firms. The three-year-old startup, which has offices in Delhi and Palo Alto, says its automated summaries are 90% accurate and can help clients save 80% of their time spent analyzing routine documents. Last year, Precily raised an undisclosed amount of funding from Pune-based VC firm Windrose Capital and Palo Alto-based law firm Inventus Law.
On the other side of the country, in Bengaluru, Akshay Deshraj and Sourabh Gupta‘s Vernacular.ai developed a voice assistant that can help companies provide quicker and better customer service. Founded in 2016, the startup’s AI technology can convert audio from telephone inquiries to text and respond in human-like conversations in more than 160 Indian dialects. Last year, the startup raised $5.1 million in financing led by Indian VC firms Exfinity Ventures and Kalaari Capital. In February, Vernacular, which has more than 100 employees, announced that it is planning to double its headcount by the year’s end.
Over in China, widely perceived as one of the global leaders in AI, Zhilin Yang is also applying AI technology to conversations. Yang cofounded Beijing-based Recurrent AI in 2016 to develop algorithms that can analyze speech. It helps monitor conversations initiated by salespeople and summarizes information such as customer preference and age group. The company’s products have been applied by several dozen Chinese companies including Zhong An Insurance and rental firm Ziroom. Recurrent AI raised in September $12 million from investors including Sequoia China and GSR Ventures. Yang holds a Ph.D. in computer science from Carnegie Mellon University.
And in Japan, which has a vision—called “Society 5.0″—of using AI to revamp its rapidly aging society, Hiroki Shimada founded an AI-powered engineer headhunting and job site. Founded in 2016, the site, called Lapras, uses AI to gather public information from databases, social media and personal blogs, without the potential candidates signing up, to automatically generate profiles to match a company’s needs. It only charges a monthly fee. Lapras estimates when that candidate might be looking to switch jobs and their skill level too. It raised ¥740 million ($7 million) so far, including most ¥350 million in September.
Source-Forbes
