Bithumb held its first in-house artificial intelligence contest and identified 107 AI use cases applied to day-to-day work, including incident response, information-leak monitoring and trading strategy verification.

Bithumb said on Oct. 26 it has wrapped up its in-house AI contest titled "AI Pro Tips, It-Thumb?"

The contest was designed to boost work efficiency by sharing employees' AI know-how and ideas used in their work processes across the organization. The contest ran from Aug. 17 to 31. A total of 110 employees took part and submitted 107 entries.

After document screening, 13 cases advanced to presentation judging. In the final presentations, seven judges selected winners based on the level of AI use and its impact on the organization.

First prize went to Joo-ho Ko (고주호), a team leader in Bithumb's IT division. Ko developed a tool that monitors in real time token usage and cache performance for the generative AI coding tool Claude Code.

Judges gave high marks to its ability to manage costs without imposing uniform limits on AI usage by selecting suitable AI models by task and optimizing cache use.

Second prize went to Seong-ryeol Eom (엄성렬), a deputy manager in Bithumb's IT division, for an "AI-based SRE incident response automation platform". The system automates repetitive tasks performed when service incidents occur, allowing staff to focus on root-cause analysis and recovery work.

Five cases deployed in actual work were selected for third prize.

They included "MCP Server centralization", which links internal knowledge and work tools into a single system, and an "AI-based information-leak monitoring platform" that analyzes tens of thousands of monthly external file-transfer records to present risk cases and the basis for judgments.

Also selected for third prize were an "advertising information pre-review helper" that provides in advance relevant laws and items to check for advertising and guidance copy, "strategy pipeline automation" that cut trading strategy verification from about 1 month to the same day level, and "Upgrade Watcher", which analyzes blockchain upgrade notices to determine the impact on deposits and withdrawals.

In addition, six cases won fourth prize, including an AI SOC Analyst and AI-based automation for vulnerability impact analysis.

Bithumb plans to use the contest as a starting point to expand the application of employees' AI use cases to actual work processes and to strengthen related infrastructure.

A Bithumb official said, "Through this contest, we confirmed that attempts to use AI can lead to positive change in planting AI DNA across the organization." The official added, "We will continue to expand AI infrastructure and steadily spread an in-house culture of innovation."