Korea Hydro & Nuclear Power (KHNP) is Korea's largest power-generation public enterprise, with about 12,000 employees nationwide. In May 2026, KHNP successfully launched an 'intelligent work communication messenger' company-wide, integrating HyperCLOVA X-based AI features into EZNIX's EzQ Messenger — one of the first messenger-AI integration cases at a public institution.

Q. What was the background for introducing AI to the messenger?

KHNP is a large organization, and smooth communication between headquarters and nationwide business sites has always been a challenge. Translation with English-speaking partners consumed considerable cost and time, while demand for internal knowledge sharing and document efficiency kept growing. We planned this project to use AI directly within the work chat window, based on the already-stable EzQ Messenger, with no need for a separate system.

Q. What AI features are built into the messenger, and which are used most?

Five major AI features were integrated

①  real-time auto-translation — instantly translate chat text with one click

② AI Assistant (HiTalk) — internal-knowledge-based Q&A and work guidance

③ document summary — summarize attachments directly in the chat window

④ conversation summary — key summaries of long conversations

⑤ Spell check — real-time checking and correction. 

Real-time auto-translation records by far the highest usage, followed by document summary, AI assistant, and spell check.

Q. Applying AI to an internal network at a security-sensitive public institution must not have been easy — how did you solve that?

Using external cloud AI services was never an option, so together with EZNIX we built the HyperCLOVA X (HCX) AI model entirely within our internal network, with four servers in a redundant setup for stability and flexible model switching via the admin console.

Q. What actual results or changes have you seen since the company-wide launch? 

The biggest change is that, through the AI assistant, employees search the work portal less and check internal knowledge directly within the messenger. Thousands of AI requests occur daily, confirming employees are naturally weaving AI into their work.

Q. Why did you choose EzQ Messenger, and what are your future plans? 

EzQ was already a proven messenger operating stably within KHNP. Adding AI to a familiar, already-used messenger was far more effective than introducing a new system. Going forward, we plan a second phase — automating IT inquiries via the employee help desk and AI, and personalized English-learning features — using the messenger as a platform.