KT has deployed a multilingual AI assistant across its South Korean retail stores, supporting over 20 languages to handle customer inquiries about rate plans, subscriptions, and service benefits without requiring fluent human staff. The assistant operates on a dedicated in-store device following a successful pilot programme, designed to triage foreign customers and reduce workload at store locations—particularly single-staff branches where language barriers previously created friction. This builds on KT's existing network of 180 "Global Stores" staffed with human multilingual consultants, which have seen foreign customer activations rise from 23 percent to 32 percent between January 2024 and February 2026. The AI deployment effectively extends language coverage from eight languages to over 20, solving a fundamental scaling problem: maintaining fluent speakers across that many languages is economically unfeasible, but a conversational terminal requires only integration and maintenance.
For CX teams, this deployment illustrates a critical shift in how support infrastructure handles geographic and linguistic fragmentation. Rather than treating multilingual support as a staffing problem, KT has reframed it as a routing and qualification problem—the AI handles initial discovery and intent-matching, freeing human agents to focus on complex transactions and relationship-building. The question for teams already managing omnichannel platforms is whether your current stack can accommodate similar triage layers without creating disconnects between AI-assisted discovery and human handoff. KT's approach also highlights a regulatory consideration often overlooked in AI deployment: identity verification and contract execution typically require human agents under telecom regulations, meaning the assistant's role is deliberately constrained to pre-sales qualification rather than full transaction completion.
The broader implication is that point-of-sale AI is becoming table stakes in markets with significant foreign customer populations or tourism. Teams building similar tools should examine whether their CX platform can integrate conversational AI terminals that feed structured customer intent data back into existing ticketing or CRM systems—otherwise you risk creating isolated AI experiences that don't reduce actual support workload. For regional carriers and retailers operating in high-immigration markets, this deployment demonstrates that AI can meaningfully improve customer experience and operational efficiency without requiring wholesale replacement of human staff, provided the AI is positioned as a qualification layer rather than a replacement.
KT deploys multilingual AI assistant in stores Let's Data Science