SuperAPI AI Directory

escalation-manager

General & other

View repository

Updated 10-15 21:35

About this skill

Escalation Manager

Overview

Handle customer complaint escalations, provide special handling for VIP customers, analyze VOC (Voice of Customer), and generate weekly CS summary reports. Reflecting 2026 Korean CX trends (feelconomy — emotions as a consumption variable), it provides emotion-based complaint response and data-driven service improvement insights. Through moai-schedules integration, automatic CS summary generation is possible every Friday at 5pm.

Trigger Keywords

escalation, customer complaint, complaint, feedback analysis, VOC, NPS, VIP handling, grievance handling, CS summary, weekly CS, customer insights

Workflow

Escalation Handling

  1. Re-assess Urgency: Check ticket-triage results or analyze directly entered content
  2. Emotion Level Analysis: Classify as anger/frustration/disappointment/neutral (reflecting feelconomy trends)
  3. Assign Escalation Level:
    • Level 1: Team leader → respond within 2 hours
    • Level 2: Department head → respond within 4 hours
    • Level 3: Executive → immediate response (service outage, media exposure risk)
  4. Check VIP Status: Based on customer tier, purchase history, social influence
  5. Establish Response Strategy: Empathize first → analyze cause → compensation plan → prevent recurrence
  6. Follow-up After Resolution: Confirmation contact upon completion, satisfaction survey

VOC Analysis and Reporting

  1. Data Collection: Collect tickets, chat logs, SNS mentions within the period
  2. Classify by Type: Complaint → improvement request → praise → inquiry
  3. Trend Analysis: Identify recurring issues, increase/decrease vs. previous week
  4. Derive Insights: Top 3 issues, root causes, improvement priorities
  5. Generate Report: Weekly CS summary markdown document

Escalation Matrix

Type Urgency Responsible Level Response Target
Service Outage Highest Level 3 (Executive) Within 30 minutes
Payment Error High Level 2 (Department Head) 1-2 hours
VIP Complaint High Level 2 (Department Head) 1-2 hours
Recurring Complaint Medium Level 1 (Team Leader) Within 4 hours
Public SNS Complaint High Level 2 + PR팀 Within 2 hours
General complaints Normal Assigned agent Within 24 hours

Weekly CS summary report format

[주간 CS 요약] {YYYY}년 {M}월 {N}주차

■ 핵심 지표
- 총 수신 티켓: N건 (전주 대비 ±N%)
- 평균 응답 시간: N시간 (목표: 4시간 이내)
- 1회 해결율(FCR): N% (목표: 80% 이상)
- 고객 만족도(CSAT): N점 / 5점

■ 유형별 현황
| 유형 | 건수 | 비율 | 전주 대비 |
|------|------|------|-----------|

■ 에스컬레이션 현황
- Level 1: N건 | Level 2: N건 | Level 3: N건

■ Top 3 이슈 & 인사이트
1. [이슈]: [원인] → [개선 권고]

■ 다음 주 액션
- [ ] ...

2026 Emotion-Based CS Response Principles

Reflecting Zendesk·Medallia 2026 CX research:

  • Feelconomy response: Empathize with emotional state (anger/frustration/disappointment) first, then present solutions
  • Human-in-the-Loop: After AI first-level classification, complex or emotional complaints must be transferred to a human agent
  • Context retention: 74% of customers complain about repeated explanations → pass previous conversation history to the escalation agent
  • AI transparency: Explicitly inform customers of what AI handled (95% of customers require this)

Deliverables

  • Escalation handling plan (level/agent/timeline)
  • VIP response special handling script
  • Weekly VOC/CS summary report
  • Top 3 recurring issue improvement recommendations
  • Customer satisfaction trend analysis

Notes

  • moai-schedules integration: Recommended to set up an automatic schedule for "weekly CS summary every Friday at 5pm"
  • Legal caution: Include handling guidance based on the Consumer Dispute Resolution Standards (Korea Fair Trade Commission) and the Act on Consumer Protection in Electronic Commerce
  • SNS crisis response: Complaints posted on public platforms (Instagram/X/Blind) follow the principle of public response within 24 hours
  • Personal information: Masking is mandatory when including customer personal information in escalation reports

Problem solving

Situation Solution
Unable to access original ticket text If you provide a summary of the complaint and the level of customer reaction, I will suggest a level and response strategy
Insufficient customer information Even without customer tier or purchase history, urgency and response direction are determined based on the inquiry content
Internal compensation policy unknown General CS compensation standards (points/coupons/refunds) are presented as examples, and policy decisions follow internal judgment
No VOC data If you provide only the list of available feedback or major complaint patterns, I will design the analysis structure

Reference materials

For the detailed guide, refer to the file below:

  • references/feedback-analyzer.md — VOC·NPS feedback analysis methodology guide

Related skills (post-processing chain)

Descriptive deliverables (escalation response plan, VIP response script, improvement recommendations, VOC insight narrative) are finalized with the following chain:

escalation-manager → moai-core:ai-slop-reviewer → moai-content:humanize-korean
  • moai-core:ai-slop-reviewer — Review and revise AI-sounding expressions, excessive modifiers, and mechanical sentence patterns
  • moai-content:humanize-korean — Polish Korean honorifics for naturalness and a human-written tone

In the weekly CS summary report, numeric·KPI tables(key metrics, status tables by type, etc.) are not subject to the post-processing chain. If a table or dashboard format is needed, moai-office:xlsx-creator or moai-data:data-visualizerroute to, and only the narrative sections of the report (Top 3 issues, insights, action recommendations) are refined through the above chain.

When not to use this skill

  • Initial ticket classification and urgency assessment → the moai-support:ticket-triage skill is more appropriate
  • Drafting general customer responses → use the moai-support:draft-response skill
  • Writing knowledge base documents → use the moai-support:kb-article skill
  • Responding to legal disputes and litigation → use the moai-legal skill and involve the legal team or a legal expert