Overview
Quickly classify and prioritize customer support tickets. In line with 2026 Korean CS standards (88% of consumers expect faster responses than the previous year), establish a system for handling within 24 hours, and after AI performs first-level classification, escalate complex/emotional complaints to human agents.
Trigger Keywords
Ticket classification, inquiry classification, CS priority, customer inquiry, support request, ticket assignment, inbound CS, inquiry handling, helpdesk
Workflow
- Ticket Receipt Analysis: Check inquiry content, channel (email/chat/phone/SNS), customer tier
- Type Classification (apply the classification system below):
- Technical issue / Payment·Refund / Shipping inquiry / General inquiry / Complaint·Complaint / VIP request
- Urgency Determination:
- P1 Urgent: Service outage, payment error, VIP complaint → immediate response (within 30 minutes)
- P2 High: Refund request, feature error → within 2-4 hours
- P3 Normal: General inquiry, usage guidance → within 24 hours
- P4 Low: Suggestions, feedback → within 48 hours
- Assignee Assignment: Connect to a specialized team by type or automated FAQ response
- Escalation Judgment: Emotional complaints, repeated inquiries, and VIPs switch to human-in-the-loop
2026 Korean CS Response Standards
Based on the Zendesk 2026 CX Trends Report:
| Standard | Figure |
|---|---|
| Expectation of 24-hour response | 74% of consumers |
| Expectation of faster responses (vs. previous year) | 88% of consumers |
| Complaints about repeated explanations during escalation | 74% of customers |
| Preference for multichannel support | 76% of consumers |
| Demand for disclosure of AI decision rationale | 95% of customers |
| Daily AI usage rate among office workers | 52% (about 2x the global average) |
Classification Output Format
[티켓 분류 결과]
티켓 ID: #XXXXX
접수 채널: 이메일 / 카카오채널 / 앱 내 문의
분류: 결제·환불
긴급도: P2 (높음)
예상 처리 시간: 2~4시간 이내
권장 담당팀: 결제팀
에스컬레이션 필요: 아니오
요약: 고객이 3일 전 주문 취소 후 환불이 미완료됨을 문의
권장 초안: cs-draft-response 스킬 호출
Deliverables
- Classification label and urgency level (P1–P4)
- Recommended team/assignee assignment
- Estimated response time calculation
- Determination of whether escalation is needed
- Suggested related KB article links (cs-kb-article skill integration)
Reference Materials
For detailed guides, refer to the file below:
references/customer-support.md— Customer Support Operations Guide
Notes
- Honorifics required: When responding to customers, using honorific language is required, and "고객님" should be used as the default form of address
- Sentiment Analysis: Tickets with a high level of complaints (peelconomy trend—emotions as a consumption variable) are automatically escalated to P1/P2
- Personal Information: When sensitive information such as resident registration numbers or card numbers is included, classify after masking
- Multimodal: For tickets with image/video attachments, summarize the content and forward the original to the assignee
Problem Resolution
| Situation | Resolution Method |
|---|---|
| Unable to access the original ticket text | Even if you enter only a summary of the inquiry or key keywords, we will guide you through the classification criteria |
| Insufficient Customer Information | Even without customer tier, urgency and type are determined based on the inquiry content |
| Conflict with Internal Classification Regulations | If you share your internal classification system, we will adjust accordingly |
| Need for Multichannel Integration | We will also design a channel-specific priority mapping table for you |
Ticket classification is a classification/priority determination task and does not generate a separate narrative deliverable. Based on the classification results, Customer Response Draftwhen drafting
moai-cs:cs-draft-responseis called, and the corresponding skill handles themoai-coworker:ai-slop-reviewer → moai-writer:korean-humanizepost-processing chain.
When Not to Use This Skill
- Drafting Customer Responses → the moai-cs:cs-draft-response skill is more appropriate
- Escalation Handling and VIP Response → use the moai-cs:cs-escalation skill
- Writing Knowledge Base Articles → use the moai-cs:cs-kb-article skill
- Operating actual ticket systems (Zendesk, Freshdesk) → you must connect directly to the relevant system