The power utility has developed an “AI-based talent recommendation system” that taps HR and job performance data to suggest candidates, rather than relying solely on traditional methods driven by human experience and instinct, according to company officials. (Image courtesy of Pixabay/CCL)
SEOUL, Mar. 18 (Korea Bizwire) – The Korea Electric Power Corporation (KEPCO), one of South Korea’s largest state-run companies, is breaking new ground by deploying artificial intelligence to assist in identifying and recommending talent for key leadership positions.
The power utility has developed an “AI-based talent recommendation system” that taps HR and job performance data to suggest candidates, rather than relying solely on traditional methods driven by human experience and instinct, according to company officials.
Previously, KEPCO made personnel decisions for major appointed roles based primarily on basic career information, internal reputations and the intuitions of those with hiring authority. But that approach had limitations in ensuring a diverse and comprehensive pool of candidates.
To address such shortcomings, the state-run corporation established a dedicated HR analytics division last year while bolstering data science capabilities at its research institute. The moves, reportedly spearheaded by President Kim Dong-cheol after his appointment in September, are part of a digital transformation push.
Engineers at the Data Science Lab developed natural language processing models capable of analyzing integrated power systems and business data. This enabled an AI-powered recommendation engine to match individuals’ skills and experiences with open job requirements.
KEPCO established a dedicated HR analytics division last year while bolstering data science capabilities at its research institute. (Image courtesy of Yonhap)
In the process, KEPCO filed four related patents covering areas like using natural language to identify talent, classifying emotions in written reviews, mapping skills to job roles, and operating the AI-based hiring system itself.
Among the innovations is an AI technology trained to classify the emotional sentiment — positive, negative or neutral — expressed in written performance reviews that incorporate narrative feedback.
For example, if an employee review states: “They are hardworking but struggle with collaboration,” the AI can detect the negative context around “collaboration” and factor that into its analysis, KEPCO explained. It can similarly parse the nuanced difference between phrases like “They are making efforts” versus “More effort is needed.”
By processing qualitative reviews with greater nuance, the AI can improve the accuracy of the “word cloud” visualizations previously used to summarize feedback, according to the company.
“Through projects like workforce planning models, evaluating the impact of promotion policies and parsing emotions in reviews, we are enhancing organizational productivity and driving internal innovation,” said a KEPCO spokesperson. The utility plans to further build data-driven digital capabilities across other operational domains, they added.
Kevin Lee (kevinlee@koreabizwire.com)
