Potatonet, a South Korean security startup, is taking on the challenge of creating a new form of defense that uses AI to visualize "malicious sites" lurking in the Deep Web and Dark Web, and prevent them before they happen. At a time when national cyber risks are on the rise, what is this technology that is attracting the keenest attention from public and financial institutions?
*This article is brought to you as part of a special feature on startups selected for the "GovTech Startup Support Program" sponsored by the Korea Software Association (KOSA) Potatonet is now seeking co-creation partners in Japan.
Preemptive defense against "invisible threats" lurking in cyberspace
As cyber attacks on a national level become a reality, we must look not only at the "surface" of the Internet, but also at the "invisible threats" that lie deep within. Security Startup from South KoreaPotatonet) is meeting this challenge head on.
Potatonet has attracted attention because of its unique AI-based technology that monitors the "Deep Web and Dark Web" and can detect and visualize malicious sites in real time. The company's founder and CEO, Kyung-Ah Shin, is a specialist in AI-based threat detection and has 30 years of experience in enhancing security in the ICT, public, and financial sectors in South Korea.
The real threats to security are those that are undetectable," she says, adding that after more than 4 years of monitoring 4.3 million websites in Korea, she realized the danger of "areas missed" by existing solutions. She realized the danger of "areas" that existing solutions "miss," and decided to found Potatonet.
AI Threat Intelligence to "Visualize" the Malicious Deep Web - "Owl Threat Intelligence".

Potatonet's core solutions are,olpemi (Olpemi/Owl Threat Intelligence), a malicious deep web detection platformIt combines an AI-based deep web analysis engine with a proprietary small language model (sLM) to provide
- Real-time analysis by AI to immediately identify dangerous websites and URLs
- Provide screenshots of malicious sites to promote "visual" understanding
- Automatic generation of attack patterns and threat context using ChatGPT3 and LLM
- Proprietary patented "hidden web" detection technology
Conventional pattern-matching detection and blacklists could not handleZero-day attacks and cloaking malwareThe design of the system has been highly evaluated. The design is highly regarded for allowing even personnel without specialized knowledge to intuitively grasp and judge risks.
Expanding use in the public sector - from PoC to full-scale introduction
Potatonet has already been introduced at the Korea Sports Promotion Agency (KSPO) and the Korea Internet Service Agency (KISA), and is currently undergoing PoC (Proof of Concept) trials at a number of public institutions and educational institutions, etc. It will record sales of approximately 340 million won (approx. 33 million yen) in 2023-2024,The highest grade 1 in the government's Software Quality Certification (GS Certification)acquired.
Furthermore, in December 2024, the company participated in the EcoSystem exchange meeting hosted by JETRO. With meetings with Bureau Veritas in Japan and the registration of the "olpemi" trademark completed, preparations for expansion in the Japanese market are steadily underway.
Undetected Attacks" Increasing in Japan - Real Needs of the Market
In 2023, Japan will see a sharp increase in ransomware and DDoS attacks originating from the Dark Web. While there are reports of information leaks at KADOKAWA and other major corporations, cyber-attacks targeting national institutions and transportation infrastructure will also continue.
In particular, Japan-related postings on the Deep Web have increased 1,000% year over year. In response to this current situation, there is a growing need from Japanese companies and local governments for preemptive measures that "catch the signs and block them.
Potatonet is already beginning to make contacts with several Japanese companies and municipalities as a rare entity that can meet these needs.
*DDoS attacks: Distributed Denial of Service attacks
Co-Creation Partners Wanted - To Protect Japan's Security Together
Potatonet envisions the following companies and organizations as partners in the Japanese market:
- Distribution company handling cyber security merchandise
→ Cooperation with distribution partners is essential, especially in the Japanese market where building long-term trust is important. - Companies supporting introduction of public offices, educational institutions, municipalities, etc.
→ Partners who can work together to enhance DX and security in the public sector - Practical departments that can help with cybersecurity PoC demonstrations
With the response already coming in from PoC and exhibitions, the company is also working on optimizing the UI, supported languages, and after-sales support system for the Japanese market, showing a positive attitude toward collaboration in a way that is suitable for the Japanese market.
Future Prospects and Messages
With the people of Japan, we must stay ahead of the "invisible threat.
Potatonet will be officially certified by the Korean government as an "innovative product" in 2025, and a provisioning system linked to the NAVER Cloud is also planned. Furthermore, we are planning to obtain CC certification (Common Criteria International) in 2026, and to deploy Potatonet as a global standard.
We will do our utmost to localize, provide ongoing support, and develop customized products in Japan.
With all companies working to strengthen security in Japan and all organizations preparing for cyber risks in the public domain,"Making invisible threats visible."We hope to co-create new solutions.
We would be happy to have the opportunity to speak with you.
Establishment: 2022 / Head office: Seoul Special City, South Korea
Representative: Shin Kyung-ah
Background: Doctor of Information Protection (AI Threat Detection), Korea University Graduate School / 30 years of experience in cyber security industry / Expert in malware response and vulnerability assessment in ICT, public and financial sectors
URL:https://potatonet.ai/






