[AI Revolution](141) Changi Airport Also Impressed... Klient Leading 'Public Procurement' Digital Transformation with AI

Jo Junho, CEO of ClienT, Interview
Holding 7 Million Bidding Data Entries
AI Analyzes Hundreds of Pages of Bidding Documents in Seconds

"The size of South Korea's public procurement market exceeds 200 trillion won. However, it still relies on manual document analysis. This means downloading and analyzing hundreds of pages of Request for Proposals (RFPs) one by one."

Jo Jun-ho, CEO of ClienT, recently explained the background of developing an AI-based bidding analysis solution in an interview with Asia Economy. Founded in 2023, ClienT provides a service that automatically analyzes vast bidding documents using AI technology and offers customized information for companies.

Cho Jun-ho, CEO of Client, is explaining the AI-based bidding analysis solution. Photo by Park Yu-jin

The biggest problem in the bidding market was the inefficient document processing process. Companies had to check every bidding announcement posted daily and analyze RFPs that could be hundreds of pages long. Especially for public institutions, since all purchases must be made through bidding, such inefficiency became a factor hindering the overall market development.

ClienT's core competitiveness lies in accumulated data and AI analysis technology. The company secured about 7 million bidding data entries from ‘Nara Market,’ the electronic procurement system operated by the Public Procurement Service. This data is analyzed using a ‘multi-agent’ system that combines OpenAI’s AI models GPT-3 and GPT-4. GPT-3 summarizes the documents, and GPT-4 extracts accurate data.

"Previously, hundreds of pages of RFPs had to be reviewed one by one. With our solution, companies can instantly check the bids they can participate in and even predict the success probability. Especially for private companies, where partner registration is often required, participation conditions can be automatically verified."

ClienT is already recognized in the global market. It established a Singapore branch last September and completed a proof of concept (PoC) for RFP generation with Changi Airport. Simply put, when Changi Airport issues projects such as robot introduction, AI drafts hundreds of pages of RFPs, enabling non-IT staff to easily review and revise them. ClienT is also collaborating with Google Singapore on document analysis.

Recently, interesting research results were also derived. Through joint research with Northwestern University’s Kellogg School of Management in the U.S., it was found that AI-related purchases by domestic public institutions have surged since 2023. He said, "Bids for GPT-4 license purchases have significantly increased, and AI is increasingly used in RFP drafting," adding, "This indicates growing demand for AI in the public sector as well."

According to CEO Jo, the biggest current issue in the AI industry is ‘efficiency improvement.’ He explained, "Reducing token costs, improving processing speed, and effectively utilizing various AI models are important tasks." In particular, ClienT is building a ‘failover’ system that automatically switches to other AI models like Google’s Gemini in case OpenAI’s service goes down.

ClienT currently offers an annual subscription service to corporate clients. Domestic bid search costs 3 million won per year, and including analysis services, it is 6 million won. The premium plan, which includes access to global bidding information, costs 10 million won annually. The company aims to raise Series A investment in the first half of this year, especially pursuing cooperation with overseas venture capital firms that have global networks.