KT Drives Financial Sector AX With an Agentic AICC

As the agentic AICC market centered on financial groups gains momentum, KT has expanded its AICC business on the back of winning a contract from Woori Bank. According to KT's announcement on the 31st, KT was selected as the operator for Woori Bank's "AI chatbot and counseling bot rebuild" project. Woori Bank's president is Jeong Jin-wan.

Following its selection, KT has begun building Woori Bank's AI-based customer consultation system. Through this project, KT has signaled a shift from delivering solution-centered offerings to expanding into platform-based business.

Conventional AICC was characterized by operation based on fixed rules and scenarios. Customer requests were also handled sequentially, creating a structure in which consultations proceeded within a predetermined flow.

By contrast, agentic AICC is distinguished by the use of a range of AI agent technologies. During consultations, it can flexibly switch tasks depending on requests and also return to work that was already in progress. It can handle complex requests and respond immediately to additional questions that arise during a task, according to the situation.

KT divides call center and consultation system build methods into 4 types based on customer requirements. Type 1 is "IPCC(IP Contact Center) deployment," aimed at rebuilding aging call infrastructure and ensuring stable operations. Type 2 is "consultation support system deployment," focused on improving consultation efficiency rather than fully introducing AI. Type 3 is "integrated AICC deployment," for the full introduction of AI-based bots and AI handling of simple tasks. Type 4 is "agent AICC deployment," which promotes AX in consultation work based on the adoption of agentic AICC.

Woori Bank's project will be carried out as the fourth type, "agent AICC deployment." The project will apply a new solution, "Agent Connector," which connects heterogeneous channels and services such as chatbots, counseling bots, Woori Bank AI banker, and AI agents. It also performs organic linkage of customer consultation context and information.

KT is applying its own STT(Speech-to-Text) and TTS(Text-to-Speech), and the LLM model is structured so that AI services are configured by selecting from a range of options as needed in collaboration with the client, said Seo Dong-cheol, who is in charge of the AX Technology Headquarters within KT's AX Business Division.

KT plans to run agentic AICC based on an agent platform. It is currently developing "Agentic On Platform" to provide AI business packages tailored to customer requirements.

Seo said the agentic platform has about 50 components and that KT developed most of the components needed to implement AI services in-house. He added that, given that client demand tends to be package-oriented, KT is preparing a range of business packages that allow users to select functions such as AICC and Agent Ops.

Under its future plans for applying agentic AICC, KT aims to expand overall work coverage from 30% to 60%, raise consultation completion rates from 75% to 80%, and cut the agent development, transition, and verification period by 50%.

Heo Yang-seok, team leader at the AX Business Headquarters within KT's AX Business Division, said existing AICC business through 2025 focused on reducing costs by improving consultation efficiency amid an insufficient AI data infrastructure environment. He added that starting in 2026, KT will pursue AI-based customer service innovation and value creation by using a variety of elements such as marketing agents and voice agents.

Source: IT DAILY · Seong Won-young
Original: https://www.itdaily.kr/news/articleView.html?idxno=241306

References

• https://www.itdaily.kr/news/articleView.html?idxno=241306