Support’s “one-time help” becomes the blueprint for company-wide AI

Why Read This from an HCI Perspective

This article is meaningful for HCI practitioners and researchers because it frames AI not as a simple automation tool, but as an interaction problem that ties together the entire customer experience. In particular, when support, sales, and success operate in silos—forcing customers to repeat the same information—that reflects a common real-world issue: “disconnected experiences.” It also prompts us to think that what matters more than where you deploy AI is how far it remembers and when humans step in.

CIT's Commentary

What’s especially interesting is that, as performance from support-focused AI is extended to other departments, “continuity of experience” emerges as the core issue rather than raw technical capability. If the sales, support, and success teams each build their own Agents, internal efficiency may improve—but from the customer’s perspective, it can feel like they’re no longer talking to the same company. That’s why this piece asks how to design shared context and history, and how to route handoffs to a human when something fails—rather than simply splitting AI Agents by function. These challenges connect directly to Korea’s domestic landscape as well, including Naver, Kakao, and startup services spanning 상담 (consulting), commerce, and subscriptions. Since channels change quickly in Korea, a conversation structure that continues in one thread may become more important than optimizing each department separately.

Questions to Consider While Reading

• Q.When you place AI Agents separately by department versus bundling them into a single shared customer Agent, what user-experience difference is most likely to show up in practice?
• Q.To prevent customers from having to re-enter the same information, what context should be shared and what information should be separated?
• Q.When the AI fails while answering, how should you design the handoff path so that customers don’t feel anxious or confused?