← 返回
未分类

客服话术专家

全场景客服话术生成器,支持电话/在线聊天/邮件/社交媒体四种服务渠道,覆盖电商/SaaS/金融/电信/旅游/医疗/教育/游戏/HR/B2B等全行业场景。提供话术生成/优化/异议处理/情绪安抚/升级管理/话术评估六大核心能力,支持中文/英文/阿拉伯语/日语四语言版本,包含完整的文化适配指南。触发词:客服话术、话术生成、优化话术、异议处理、客户生气、升级处理、评价话术。
全场景客服话术生成器,支持电话/在线聊天/邮件/社交媒体四种服务渠道,覆盖电商/SaaS/金融/电信/旅游/医疗/教育/游戏/HR/B2B等全行业场景。提供话术生成/优化/异议处理/情绪安抚/升级管理/话术评估六大核心能力,支持中文/英文/阿拉伯语/日语四语言版本,包含完整的文化适配指南。触发词:客服话术、话术生成、优化话术、异议处理、客户生气、升级处理、评价话术。
user_3ef5c419
未分类 community v1.0.0 1 版本 66666.7 Key: 无需
★ 0
Stars
📥 2
下载
💾 0
安装
1
版本
#latest

概述

客服话术专家 (CS Script Expert)

全场景客服话术生成、优化与评估工具。

Overview

This skill enables generation of professional customer service scripts for phone, live chat, email, and social media interactions. It covers script generation, optimization, objection handling, emotional de-escalation, escalation management, and script evaluation. The skill supports Chinese, English, Arabic, and Japanese with appropriate cultural adaptations.


When to Use This Skill

Trigger this skill when users request any of the following:

  • Script generation: "帮我写一段客服话术"、"Generate a customer service script for..."
  • Script optimization: "帮我优化这段话术"、"How can I improve this script..."
  • Objection handling: "客户说太贵了怎么回应"、"How to handle 'it's too expensive'..."
  • Emotional de-escalation: "客户很生气怎么办"、"Angry customer script..."
  • Escalation management: "什么时候应该升级"、"When to escalate to manager..."
  • Script evaluation: "评价这段话术"、"Evaluate this script quality..."

Decision Tree

User request
├── Generate new script → See "Script Generation" section
├── Optimize existing script → See "Script Optimization" section
├── Handle objection → See "Objection Handling" section
├── De-escalate emotion → See "Emotional De-escalation" section
├── Manage escalation → See "Escalation Handling" section
└── Evaluate script → See "Script Evaluation" section

Script Generation

Step 1: Gather Context

Before generating a script, identify these key parameters:

  1. Service channel (required):
    • Phone (电话)
    • Live chat (在线聊天)
    • Email (邮件)
    • Social media (社交媒体)
  1. Industry (recommended):
    • E-commerce (电商)
    • SaaS/Software (软件)
    • Financial services (金融)
    • Telecom (电信)
    • Travel/Hospitality (旅游/酒店)
    • Healthcare (医疗)
    • Education (教育)
    • Gaming (游戏)
    • HR/Recruiting (人力资源)
    • General B2B (企业服务)
    • If not specified, use generic framework
  1. Scenario (required):
    • What is the customer's issue or request?
    • What stage of the interaction?
  1. Style (optional, default: friendly):
    • Professional (专业严谨)
    • Friendly (亲切友好) - default
    • Efficient (高效直接)
  1. Language (required):
    • Chinese (中文)
    • English (英文)
    • Arabic (阿拉伯语)
    • Japanese (日语)

Step 2: Load Relevant References

Based on the context, load the appropriate reference files:

  • Always load: references/script-frameworks.md (core frameworks)
  • If industry specified: references/industry-patterns.md (industry-specific patterns)
  • If language is non-Chinese: Load corresponding language guide (references/[language]-guide.md)

Step 3: Generate Script

Using the frameworks and patterns from references, generate a complete script that includes:

  1. Opening - Greeting and establishing connection
  2. Empathy/Acknowledgment - Making customer feel understood
  3. Problem Clarification - Confirming understanding
  4. Solution - Providing clear action steps
  5. Closing - Confirming resolution and offering further help

Step 4: Quality Check

Before presenting the script, verify:

  • [ ] Appropriate empathy expressed
  • [ ] Solution is specific with timeline
  • [ ] Language/culture appropriate for target language
  • [ ] No ambiguous commitments
  • [ ] Clear next steps provided
  • [ ] Professional yet warm tone maintained

Step 5: Present Script

Present the script in a clear format:

# [Scenario Name] - [Channel] - [Language]

## Script

[Full script here]

## Key Points
- [Important note 1]
- [Important note 2]

## Alternative Phrases
- Option A: [phrase]
- Option B: [phrase]

Script Optimization

When User Provides Existing Script

  1. Load evaluation framework: references/evaluation-framework.md
  2. Evaluate script using the 5-dimension framework:
    • Empathy & Emotional Connection (25%)
    • Problem-Solving Ability (25%)
    • Clarity & Readability (20%)
    • Professionalism & Accuracy (15%)
    • Action-Oriented & Closed Loop (15%)
  3. Provide optimized version with explanation of improvements
  4. Present in evaluation report format (see references/evaluation-framework.md section 4.1)

Objection Handling

When User Faces Customer Objection

  1. Identify objection type:
    • Price objection ("too expensive")
    • Time objection ("too slow")
    • Quality objection ("poor quality")
    • Trust objection ("don't believe you")
    • Decision delay ("need to think")
    • Service attitude objection ("bad attitude")
  1. Load objection framework: references/objection-escalation.md section 1.2
  1. Apply LAARC model:
    • Listen - Let customer express fully
    • Acknowledge - Validate feelings (not necessarily agreement)
    • Address - Provide facts/solutions/alternatives
    • Check - Confirm if customer accepts
  1. Generate response script using the framework for the specific objection type

Emotional De-escalation

When Customer is Angry/Upset

  1. Load de-escalation techniques: references/objection-escalation.md section 2
  1. Follow steps:
    • Step 1: Let customer vent (don't interrupt)
    • Step 2: Acknowledge emotions sincerely
    • Step 3: Apologize (for experience, not necessarily fault)
    • Step 4: Solve problem with options
    • Step 5: Follow up to confirm resolution
  1. Generate de-escalation script appropriate for channel and language

Escalation Handling

When to Escalate (see references/objection-escalation.md section 3.1)

Must escalate:

  • Customer explicitly requests manager
  • Legal threats or regulatory complaints
  • Amount exceeds authority
  • Technical issue can't be resolved normally
  • Customer extremely emotional, can't de-escalate
  • Cross-department coordination needed urgently

Consider escalation:

  • Uncertain about solution
  • Recurring problem needs systemic fix

Should NOT escalate:

  • Customer habitually says "I want your leader" but can be pacified
  • Problem within authority, customer just impatient
  • Avoiding difficult conversation

Escalation Script Generation

Generate scripts for:

  1. Customer requests manager - How to handle gracefully
  2. You initiate escalation - How to explain to customer
  3. Handover to next person - What info to transfer

See references/objection-escalation.md section 3.3 for templates.


Script Evaluation

When User Asks to Evaluate Script

  1. Load evaluation framework: references/evaluation-framework.md
  1. Apply 5-dimension scoring:
    • Empathy & Emotional Connection: X/10
    • Problem-Solving Ability: X/10
    • Clarity & Readability: X/10
    • Professionalism & Accuracy: X/10
    • Action-Oriented & Closed Loop: X/10
  1. Calculate total score: Weighted average
  1. Provide evaluation report in format:
    • Total score and grade
    • Scores by dimension
    • Strengths (1-3 points)
    • Areas for improvement (prioritized)
    • Optimization suggestions
    • Optimized script version
    • Explanation of improvements

See references/evaluation-framework.md section 4.1 for full template.


Multilingual Guidelines

Chinese (中文)

  • Default style: Friendly (亲切友好)
  • Key considerations:
  • Use "您" not "你" (unless brand allows "你")
  • Avoid overusing "亲" (Taobao-style, unprofessional in formal contexts)
  • Moderation with emojis and tone words
  • Load: references/chinese-guide.md

English

  • Default style: Friendly-professional
  • Key considerations:
  • Direct but polite (don't over-apologize)
  • Use contractions for warmth ("I'm", "we'll")
  • Provide options, not ultimatums
  • Regional variations: US (casual), UK (reserved), Australia (casual-respectful), India (formal)
  • Load: references/english-guide.md

Arabic (العربية)

  • Default style: Respectful (respectful/hospitality-focused)
  • Key considerations:
  • Use appropriate Islamic greetings (especially during Ramadan)
  • Respect prayer times
  • Use standard Arabic (العربية الفصحى) if unsure about dialect
  • Gender considerations in conservative regions
  • Load: references/arabic-guide.md

Japanese (日本語)

  • Default style: Highly formal (keigo mandatory)
  • Key considerations:
  • Keigo (敬語) is MANDATORY - use 丁寧語, 尊敬語, 謙譲語 appropriately
  • Customer is "お客様" (never "あなた")
  • Apologize FIRST, then discuss solution (critical in Japanese CS)
  • Never say "できません" directly - use "致しかねます" + alternative
  • Load: references/japanese-guide.md

Resources

This skill includes the following reference files:

Core Frameworks

  • references/script-frameworks.md - Script structures by channel (phone/chat/email/social), style guides, quality checklist
  • references/industry-patterns.md - Industry-specific scenarios and considerations (e-commerce, SaaS, finance, telecom, travel, healthcare, education, gaming, HR, B2B)
  • references/evaluation-framework.md - 5-dimension evaluation framework, scoring system, optimization suggestions

Objection & Escalation

  • references/objection-escalation.md - Objection handling (LAARC model), emotional de-escalation techniques, escalation guidelines and scripts

Language Guides

  • references/chinese-guide.md - Chinese language and cultural guide
  • references/english-guide.md - English language and cultural guide
  • references/arabic-guide.md - Arabic language and cultural guide
  • references/japanese-guide.md - Japanese language and cultural guide (keigo focus)

When to Load Which Resources

User request typeLoad these references
----------------------------------------
Generate script (general)script-frameworks.md
Generate script (specific industry)script-frameworks.md + industry-patterns.md
Generate script (non-Chinese)script-frameworks.md + [language]-guide.md
Optimize scriptevaluation-framework.md
Handle objectionobjection-escalation.md
De-escalate emotionobjection-escalation.md
Escalation managementobjection-escalation.md
Evaluate scriptevaluation-framework.md

Important Notes

  1. Framework-based, not exhaustive - This skill provides frameworks and patterns, not an exhaustive list of all scenarios. Use the frameworks to handle new/unseen scenarios.
  1. Brand voice overrides - If user provides brand guidelines, prioritize those over generic frameworks in this skill.
  1. Cultural sensitivity - Always consider cultural appropriateness for the target language/market. When in doubt, ask user about cultural preferences.
  1. No external data dependency - This skill is fully self-contained. No external APIs, databases, or knowledge bases required.
  1. Continuous improvement - If you discover better frameworks or examples through usage, note them for potential skill improvement.

版本历史

共 1 个版本

  • v1.0.0 Initial release 当前
    2026-06-11 17:19 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

self-improving agent

pskoett
捕获经验教训、错误和纠正,以实现持续改进。使用时机:(1)命令或操作意外失败;(2)用户纠正……
★ 4,055 📥 795,672
ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,349 📥 317,677
security-compliance

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,210 📥 266,127