SEOUL, Feb. 17 (Korea Bizwire) — When the latest version of ChatGPT 3.5 was launched by Open AI late last year, a new battle for artificial intelligence chatbots heated up across the world.
While tech giants and startups alike are dreaming of forming the industry and reshaping what people want from search features, Microsoft Corp. and Google swiftly came up with their respective chatbots, Bing and Bard, claiming their products will be the future of AI search armed with detailed responses and fluid sentences in a colloquial style.
South Korea’s big IT firms like Naver Corp. and Kakao Corp. have also announced plans to introduce their own versions of AI search services in the coming future.
Amid the back-to-back announcements by big names, the South Korean startup Allganize Inc. recently released “Alli GPT,” its signature search service for companies based on the latest ChatGPT’s open codes.
Alli is a corporate search chatbot that intelligently responds to customers through natural chat conversation flows, helping workers save time to look for a certain document or information from piles of reports, written manuals and scripts stored in a company’s database and improve their productivity.
“We provide natural language AI search solutions to companies,” Shin Kibin, chief AI officer and co-founder of Allganize, which was founded in 2017 in California, said in an interview with Yonhap News Agency last week.
“Natural language AI search finds out the exact answer right away. Normal search services based on keywords give you a list of the answers, and you have to click the page and choose the right one,” he explained.
“For example, if you want to know the inaugural winner of the Turing Awards, search engines presents you threads of links. But natural language AI search tools give you the answer directly: Alan Perlis.”
He said Alli uses the natural language model and works almost the same way that ChatGPT does.
After a company uploads its files of reports, personnel records and documents to Alli’s server, Alli recognizes all of the files in a comprehensive way, studies them itself and responds to company workers’ queries with smooth sentences with reference papers.
“If you ask Alli for the phone number of Mr. Song in your company, Alli shows you the number directly instead of showing lists of his other information,” Shin said.
The official said Allganize converts various kinds of corporate data in the forms of Microsoft Word, Excel and PDF files into computing language, and helps Alli study them in advance.
“Many think AI projects to improve productivity in a corporation are very hard tasks, because they have to prepare massive amounts of data and help the AI program learn the data. It’s very expensive and time-consuming,” he said.
“Alli takes over that duty.”
He said Alli has been welcomed by some 2,500 companies at home and abroad, including South Korea’s SK Chemicals Co., Hyundai Card Co. and KB Securities Co., and Japan’s SMBC Group, KDDI Corp. and Nitori Co.
In particular, the Japanese multinational banking and financial service firm SMBC Group became Allganize’s customer in 2019 after Allganize’s Alli outperformed IBM’s corporate chatbot program.
“IBM’s program reached an accuracy rate of 91 percent after six months, but Alli got 95 percent after two weeks,” he said. “Since then, SMBC changed its intranet search program to Alli.”
For its overseas customers, Allganize has four offices: one each in Seoul, San Francisco, Houston and Tokyo, to which the company moved its headquarters last year in a bid to go public on a Japanese stock exchange in 2025.
Shin, a graduate of the national Korea Advanced Institute of Science and Technology (KAIST) and an employee at Allganize after working at Naver and LINE Plus Corp., said its technology is aimed at helping companies lead innovation for a small cost through up-to-date AI technology.
“ChatGP became a sensation not because it mimics a human conversationalist but because it makes an inference. It shows signs of a computer that is able to lead dialogue and think,” he said.
“Allganize uses AI to help innovate in a faster way for a small cost and little effort.”