# Send Customer Support Operations Engine to your agent
Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.
## Fast path
- Download the package from Yavira.
- Extract it into a folder your agent can access.
- Paste one of the prompts below and point your agent at the extracted folder.
## Suggested prompts
### New install

```text
I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete.
```
### Upgrade existing

```text
I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run.
```
## Machine-readable fields
```json
{
  "schemaVersion": "1.0",
  "item": {
    "slug": "afrexai-support-operations",
    "name": "Customer Support Operations Engine",
    "source": "tencent",
    "type": "skill",
    "category": "效率提升",
    "sourceUrl": "https://clawhub.ai/1kalin/afrexai-support-operations",
    "canonicalUrl": "https://clawhub.ai/1kalin/afrexai-support-operations",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadUrl": "/downloads/afrexai-support-operations",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-support-operations",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "packageFormat": "ZIP package",
    "primaryDoc": "SKILL.md",
    "includedAssets": [
      "README.md",
      "SKILL.md"
    ],
    "downloadMode": "redirect",
    "sourceHealth": {
      "source": "tencent",
      "slug": "afrexai-support-operations",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-29T15:43:52.161Z",
      "expiresAt": "2026-05-06T15:43:52.161Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-support-operations",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-support-operations",
        "contentDisposition": "attachment; filename=\"afrexai-support-operations-1.0.0.zip\"",
        "redirectLocation": null,
        "bodySnippet": null,
        "slug": "afrexai-support-operations"
      },
      "scope": "item",
      "summary": "Item download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this item.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/afrexai-support-operations"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    }
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/afrexai-support-operations",
    "downloadUrl": "https://openagent3.xyz/downloads/afrexai-support-operations",
    "agentUrl": "https://openagent3.xyz/skills/afrexai-support-operations/agent",
    "manifestUrl": "https://openagent3.xyz/skills/afrexai-support-operations/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/afrexai-support-operations/agent.md"
  }
}
```
## Documentation

### Customer Support Operations Engine

You are a customer support operations architect. Help the user build, optimize, and scale their entire support function — from first ticket to mature, multi-channel, data-driven support organization.

### Phase 1 — Support Function Assessment

Before optimizing, understand current state.

### Quick Health Triage

Signal🔴 Critical🟡 Warning🟢 HealthyFirst Response Time>24h4-24h<4hResolution Time>72h24-72h<24hCSAT Score<70%70-85%>85%First Contact Resolution<50%50-70%>70%Ticket Backlog>3x daily volume1-3x<1x dailyAgent Utilization>90% or <40%40-60% or 80-90%60-80%Escalation Rate>30%15-30%<15%Customer Effort Score>4 (high effort)3-4<3 (low effort)

### Support Assessment Brief

support_assessment:
  company: "[Company Name]"
  product_type: "[SaaS/E-commerce/Marketplace/Hardware/Service]"
  date: "YYYY-MM-DD"
  
  current_state:
    team_size: 0
    channels: []  # email, chat, phone, social, in-app
    tools: []  # helpdesk, CRM, knowledge base
    monthly_ticket_volume: 0
    avg_first_response_time: ""
    avg_resolution_time: ""
    csat_score: 0
    fcr_rate: 0
    
  top_issues:
    - category: ""
      percentage: 0
      typical_resolution: ""
    - category: ""
      percentage: 0
      typical_resolution: ""
      
  pain_points: []
  goals: []
  budget_constraints: ""

### Channel Selection Matrix

ChannelBest ForResponse ExpectationCost/TicketComplexityEmail/TicketComplex issues, documentation trail4-24h$$LowLive ChatQuick questions, browsing support<2 min$$$MediumPhoneUrgent issues, complex explanationsImmediate$$$$HighSelf-Service/KBCommon questions, how-tosInstant$Medium (setup)In-AppContextual help, onboarding<5 min$$MediumSocial MediaPublic issues, brand monitoring<1h$$MediumCommunity ForumPeer support, feature discussion4-24h$LowChatbot/AIL0 deflection, routing, FAQInstant$High (setup)

### Channel Architecture by Company Stage

Startup (0-1K tickets/mo):

Email + Knowledge Base + In-App chat
1-3 agents, everyone does everything
Tool: Intercom, Freshdesk, or Help Scout

Growth (1K-10K tickets/mo):

Add: Live chat + Phone (for enterprise) + Chatbot
Tiered team (L1/L2), dedicated KB manager
Tool: Zendesk, Intercom, or Freshdesk

Scale (10K+ tickets/mo):

All channels + AI deflection + Community
Specialized teams by channel/product/tier
Tool: Zendesk Suite, Salesforce Service Cloud

### Channel Routing Logic

INCOMING TICKET:
├── Is it from a VIP/Enterprise customer?
│   └── YES → Priority queue → Senior agent
├── Can AI/bot answer with >90% confidence?
│   └── YES → Auto-respond → Offer human escalation
├── Is it a known issue with existing solution?
│   └── YES → Auto-suggest KB article → Close if confirmed
├── Complexity assessment:
│   ├── Simple (how-to, password reset, billing) → L1
│   ├── Technical (bug, integration, API) → L2
│   └── Critical (outage, data loss, security) → L2 + escalation
└── Channel-specific routing:
    ├── Social → Social team (public response <1h)
    ├── Phone → Available phone agent (no queue >3 min)
    └── Email/Chat → Round-robin by skill match

### Ticket Lifecycle

NEW → OPEN → PENDING → SOLVED → CLOSED
         ↓        ↑
      ESCALATED ──┘

Stage Definitions:

StageOwnerMax TimeExit CriteriaNewUnassigned15 minAgent picks up or auto-assignedOpenAgentVaries by priorityWorking on resolutionPendingCustomer72h auto-close warningWaiting for customer responseEscalatedL2/Specialist4h acknowledgmentNeeds specialist knowledgeSolvedAgent48h auto-closeSolution provided, awaiting confirmationClosedSystem—Confirmed resolved or auto-closed

### Priority Matrix

PriorityCriteriaFirst ResponseResolution TargetP0 — CriticalService down, data loss, security breach15 min4hP1 — HighMajor feature broken, revenue impact1h8hP2 — NormalFeature issue, workaround exists4h24hP3 — LowHow-to, enhancement request, cosmetic24h72h

### Auto-Priority Rules

auto_priority:
  P0_triggers:
    - keyword_match: ["outage", "down", "data loss", "breach", "can't login all"]
    - customer_tier: "enterprise"
    - affected_users: ">100"
    
  P1_triggers:
    - keyword_match: ["broken", "not working", "error", "billing issue"]
    - customer_tier: "business"
    - revenue_impact: true
    
  P2_default: true  # Everything else starts here
  
  P3_triggers:
    - keyword_match: ["feature request", "nice to have", "suggestion"]
    - category: "enhancement"

### Ticket Quality Checklist

Every ticket response should include:

Greeting — personalized, warm, matches tone
 Acknowledgment — restate the issue to confirm understanding
 Resolution/Next Step — clear action taken or planned
 Timeline — when they can expect resolution if not immediate
 Prevention — how to avoid this in future (when applicable)
 Closing — invitation to reach out again, satisfaction check

### Ticket Tags & Categories

taxonomy:
  categories:
    - account: [login, password, billing, subscription, permissions]
    - product: [bug, feature_request, how_to, integration, performance]
    - onboarding: [setup, migration, training, documentation]
    - technical: [api, webhook, sso, data_export, custom_config]
    - feedback: [complaint, compliment, suggestion, survey_response]
    
  sentiment: [positive, neutral, negative, urgent]
  
  root_cause:
    - user_error
    - documentation_gap
    - product_bug
    - missing_feature
    - third_party_issue
    - billing_system
    
  resolution_type:
    - self_service_redirect
    - agent_resolved
    - engineering_fix
    - product_change
    - refund_credit
    - no_action_needed

### The HEART Response Method

Every customer interaction follows HEART:

Hear — Read the full message. Understand the real problem, not just the stated one.
Empathize — Acknowledge their frustration. Validate the experience.
Act — Take concrete action. Explain what you're doing.
Resolve — Provide the solution or clear next steps with timeline.
Thank — Thank them for reaching out. Confirm they're satisfied.

### Response Templates

Template 1: Bug Report Acknowledgment

Hi [Name],

Thanks for reporting this — I can see how [specific impact] would be frustrating.

I've reproduced the issue on my end and confirmed [what you found]. I'm escalating this to our engineering team with priority [P level].

Here's what happens next:
- Engineering will investigate within [timeframe]
- I'll update you as soon as we have a fix or workaround
- In the meantime, you can [workaround if available]

Reference: [Ticket #]

Let me know if anything changes on your end. I'm on this until it's resolved.

[Agent Name]

Template 2: Feature Request Response

Hi [Name],

Great suggestion — [specific feature] would definitely [acknowledge the value].

I've logged this as a feature request and linked it to [X] similar requests from other customers. Our product team reviews these monthly to prioritize the roadmap.

While I can't promise a timeline, the volume of requests for this is helping make the case. I'll tag you on any updates.

In the meantime, have you tried [alternative approach]? It's not exactly what you're after, but some customers find it helpful for [use case].

Thanks for taking the time to share this — feedback like yours directly shapes what we build.

[Agent Name]

Template 3: Angry Customer De-escalation

Hi [Name],

I hear you, and I'm sorry — this isn't the experience you should be having with [product].

Let me be direct about what happened: [honest explanation without excuses].

Here's what I'm doing right now:
1. [Immediate action]
2. [Next step with timeline]
3. [Compensation/goodwill if appropriate]

I take full ownership of getting this resolved. You'll hear from me by [specific time], not with an update that we're "still working on it" — with an actual resolution.

[Agent Name]

Template 4: Billing Issue Resolution

Hi [Name],

I've looked into your billing concern and here's what I found:

[Clear explanation of what happened with the charge]

Action taken: [refund processed / credit applied / correction made]
- Amount: $[X]
- You'll see this reflected within [timeframe]
- Reference: [transaction ID]

To prevent this going forward: [what changed or what to watch for].

Everything look right? Happy to walk through your billing history if you'd like a full review.

[Agent Name]

Template 5: Saying No Gracefully

Hi [Name],

I understand why you'd want [requested action] — it makes sense given [their situation].

Unfortunately, I'm not able to [specific thing] because [honest reason — not "our policy says"].

Here's what I can do instead:
- Option A: [alternative that partially addresses their need]
- Option B: [different approach]
- Option C: [escalation path if they want to pursue further]

Which of these works best for you? Or if none of these hit the mark, let me know what you're ultimately trying to achieve and I'll see what else we can figure out.

[Agent Name]

### Tone Calibration Guide

Customer ToneMatch WithExample ShiftCasual/FriendlyWarm, conversational"Hey! Let me take a look..."Professional/FormalClear, structured"Thank you for contacting us. I've reviewed..."Frustrated/AngryCalm, empathetic, action-oriented"I understand. Let me fix this right now."Technical/DetailedPrecise, detailed, technical"The API returns 429 when..."Confused/LostSimple, step-by-step"No worries! Here's exactly what to do..."

### Response Quality Scoring (0-100)

DimensionWeightCriteriaAccuracy25%Correct information, proper diagnosis, right solutionEmpathy20%Acknowledged feelings, personalized, human toneCompleteness20%Addressed all questions, proactive info, prevention tipsClarity15%Easy to follow, no jargon, proper formattingEfficiency10%Resolved in minimum exchanges, no unnecessary back-and-forthBrand Voice10%Consistent tone, matches company personality

Scoring:

90-100: Exceptional — use as training example
70-89: Good — meets standards
50-69: Needs improvement — coaching required
Below 50: Failed — requires retraining

### Support Tier Architecture

L0 — Self-Service / AI
├── Knowledge base, chatbot, automated responses
├── Target: Deflect 30-50% of inbound volume
└── Escalates to L1 when: confidence <90%, customer requests human

L1 — Front-Line Support
├── Common issues, account management, how-to
├── Skills: Product knowledge, communication, troubleshooting basics
├── Metrics: FCR >70%, CSAT >85%, AHT <15 min
└── Escalates to L2 when: technical depth needed, bug confirmed, >30 min

L2 — Technical / Specialist Support  
├── Complex bugs, API issues, integrations, data problems
├── Skills: Technical debugging, log analysis, API knowledge
├── Metrics: Resolution <24h, CSAT >90%, escalation to eng <20%
└── Escalates to Engineering when: code fix needed, infra issue

L3 — Engineering Support
├── Production bugs, infrastructure issues, security
├── Skills: Code access, deployment ability, database access
├── Metrics: MTTR, change failure rate
└── Escalates to Management when: customer impact >threshold

### Escalation Decision Matrix

TriggerActionTimelineP0 incidentImmediate L2 + Engineering + Manager notification15 minCustomer threatens churn (ARR >$10K)L2 + Account Manager + CS lead1hLegal threat or compliance issueL2 + Legal + Manager1hSame issue reported 3+ times by customerL2 + Bug report + PM notification4hAgent stuck >30 min on single ticketL2 peer assist or escalation30 minCustomer requests managerTransfer to team lead — never refuseImmediateSocial media escalation (public)Social team + PR if viral risk30 min

### Escalation Handoff Template

escalation:
  ticket_id: ""
  customer:
    name: ""
    tier: ""  # free/pro/enterprise
    arr: 0
    sentiment: ""  # frustrated/angry/neutral
    previous_escalations: 0
    
  issue:
    summary: ""
    category: ""
    priority: ""
    started: "YYYY-MM-DD HH:MM"
    
  what_tried:
    - action: ""
      result: ""
    - action: ""
      result: ""
      
  what_needed: ""
  customer_expectation: ""
  urgency_reason: ""

### KB Architecture

Knowledge Base
├── Getting Started (onboarding flow)
│   ├── Quick start guide
│   ├── Account setup
│   └── First [key action]
├── How-To Guides (task-based)
│   ├── By feature area
│   └── By user role
├── Troubleshooting (problem-based)
│   ├── Common errors
│   ├── Known issues
│   └── Diagnostic steps
├── API / Developer Docs (technical)
│   ├── Authentication
│   ├── Endpoints
│   └── Webhooks
├── Billing & Account
│   ├── Plans & pricing
│   ├── Payment methods
│   └── Invoices & receipts
└── FAQ (curated top questions)

### Article Quality Checklist

Title is a question or action phrase (not a label)
 First paragraph answers the question directly
 Steps are numbered, specific, and testable
 Screenshots or GIFs for visual steps (annotated)
 Edge cases covered (what if X doesn't work?)
 Related articles linked at bottom
 Last updated date visible
 Feedback widget enabled (Was this helpful?)
 SEO — title matches how customers would search
 Reading level — Grade 8 or below (Hemingway test)

### Self-Service Deflection Strategy

Target: 30-50% ticket deflection through self-service

MethodExpected DeflectionSetup EffortContextual help (in-app tooltips)10-15%MediumSearch-optimized KB15-25%MediumAI chatbot (FAQ + KB search)10-20%HighGuided troubleshooting flows5-10%MediumCommunity forum (peer support)5-10%LowVideo tutorials3-5%High

### KB Maintenance Cadence

FrequencyActionWeeklyReview "Was this helpful? No" feedback, fix top offendersMonthlyAudit top 20 search queries — ensure articles exist for eachMonthlyReview 0-view articles — update, redirect, or archiveQuarterlyFull KB audit — freshness check, accuracy reviewPer releaseUpdate affected articles before feature ships

### Content Gap Detection

FOR EACH top support ticket category:
  1. Search KB for matching article
  2. IF no article exists → CREATE (priority = ticket volume)
  3. IF article exists but tickets persist → IMPROVE (unclear or incomplete)
  4. IF article exists and is good → check discoverability (search, in-app links)

### Core Metrics Dashboard

weekly_dashboard:
  date_range: "YYYY-MM-DD to YYYY-MM-DD"
  
  volume:
    total_tickets: 0
    new_tickets: 0
    resolved_tickets: 0
    backlog: 0
    tickets_per_agent: 0
    
  speed:
    avg_first_response_time: ""
    median_first_response_time: ""
    avg_resolution_time: ""
    p95_resolution_time: ""
    
  quality:
    csat_score: 0  # target: >85%
    fcr_rate: 0  # target: >70%
    customer_effort_score: 0  # target: <3
    nps_from_support: 0  # target: >40
    
  efficiency:
    cost_per_ticket: 0
    tickets_per_agent_per_day: 0  # healthy: 15-25
    self_service_deflection_rate: 0  # target: >30%
    automation_rate: 0
    
  team:
    agent_satisfaction: 0
    attrition_rate: 0  # annual, target: <25%
    avg_handle_time: ""
    utilization: 0  # target: 60-80%
    
  trends:
    ticket_volume_wow: ""  # +X% or -X%
    csat_trend: ""
    top_issue_changes: []

### Metric Benchmarks by Company Stage

MetricStartupGrowthScaleWorld-ClassFirst Response (email)<24h<4h<1h<15 minFirst Response (chat)<5 min<2 min<1 min<30 secCSAT>75%>80%>85%>90%FCR>50%>65%>75%>85%Self-Service Deflection>10%>25%>40%>60%Cost per TicketN/A<$25<$15<$8Agent Utilization40-90%60-80%65-80%70-80%

### Root Cause Analysis

Run monthly — categorize ALL tickets by root cause:

root_cause_analysis:
  month: "YYYY-MM"
  total_tickets: 0
  
  categories:
    - cause: "Documentation gap"
      count: 0
      percentage: 0
      action: "Create/update KB articles"
      owner: ""
      
    - cause: "Product bug"
      count: 0
      percentage: 0
      action: "File engineering tickets, prioritize by volume"
      owner: ""
      
    - cause: "UX confusion"
      count: 0
      percentage: 0
      action: "Share with product/design for improvement"
      owner: ""
      
    - cause: "Missing feature"
      count: 0
      percentage: 0
      action: "Aggregate for product roadmap input"
      owner: ""
      
    - cause: "User error (despite good docs)"
      count: 0
      percentage: 0
      action: "In-app guidance, onboarding improvement"
      owner: ""
      
    - cause: "Third-party/integration issue"
      count: 0
      percentage: 0
      action: "Partner communication, status page"
      owner: ""
      
    - cause: "Billing/account"
      count: 0
      percentage: 0
      action: "Process automation, self-service billing"
      owner: ""

The 10x Rule: Every bug that generates >10 tickets/month should be escalated to engineering as a P1 fix. Every question asked >20 times/month should have a KB article AND in-app guidance.

### Team Sizing Formula

Required agents = (Monthly tickets × Avg handle time in hours) / 
                  (Working hours per agent per month × Target utilization)

Example:
- 5,000 tickets/mo × 0.25h avg handle = 1,250 hours needed
- 160 hours/agent/mo × 0.75 utilization = 120 productive hours/agent
- 1,250 / 120 = ~11 agents needed

Add buffer for:

PTO/sick leave: +15%
Training time: +10%
Peak periods: +20%
Growth: +10% per quarter

### Team Structure by Size

1-3 Agents (Startup):

Everyone is generalist
Shared queue, no tiers
Manager = agent + admin

4-10 Agents (Growth):

L1/L2 split
Team lead (50% tickets, 50% coaching)
KB owner (shared responsibility)
Specialization by product area begins

11-30 Agents (Scale):

L1/L2/L3 tiers
Dedicated team leads (1:6-8 ratio)
KB/self-service team
Quality assurance reviewer
Workforce management
Support ops/tooling

30+ Agents (Enterprise):

All above + regional teams
Dedicated training team
Support engineering team
Customer advocacy/VoC role
Director + managers hierarchy

### Hiring Scorecard

DimensionWeightWhat to AssessCommunication30%Writing clarity, empathy, tone matchingProblem-Solving25%Diagnostic thinking, creative solutionsTechnical Aptitude20%Learning speed, comfort with toolsEmotional Intelligence15%Handling frustration, de-escalationCultural Fit10%Team collaboration, growth mindset

### Interview: Support Simulation Exercise

Give candidates a real (anonymized) ticket and ask them to:

Write a response (assess communication + accuracy)
Explain their diagnosis process (assess problem-solving)
Role-play an angry customer call (assess EQ + de-escalation)
Navigate your helpdesk tool (assess technical aptitude)

Score each 1-5. Minimum 3.5 average to hire.

### Agent Onboarding Checklist (First 30 Days)

Week 1: Foundation

Product walkthrough (become a power user)
 Tool training (helpdesk, KB, CRM)
 Shadow 20+ tickets with senior agent
 Read top 50 KB articles
 Practice responses with templates

Week 2: Guided Practice

Handle L1 tickets with mentor review
 Complete 10 supervised responses
 Learn escalation procedures
 Study top 10 issue categories
 Pass product knowledge quiz (>80%)

Week 3-4: Independent with Safety Net

Handle L1 queue independently
 QA review on 50% of tickets
 First 1:1 with team lead
 Set 30-day performance goals
 Identify personal development areas

### QA Review Framework

Review cadence:

New agents (0-90 days): 30% of tickets reviewed
Experienced agents: 10% of tickets reviewed (random sample)
All escalated tickets: 100% reviewed
All negative CSAT: 100% reviewed

### QA Scorecard (per ticket)

CategoryPointsCriteriaAccuracy/25Correct diagnosis, right solution, no misinformationCommunication/25Clear, empathetic, professional, matched toneProcess/20Proper tags, priority, escalation if needed, notesEfficiency/15Minimum touches to resolve, no unnecessary delaysGoing Above/15Proactive help, prevention tips, personal touchTotal/100

Score thresholds:

90+: Exceptional — recognition, potential mentor
75-89: Meets expectations
60-74: Coaching needed — create improvement plan
Below 60: Performance concern — immediate coaching + daily review

### QA Calibration Sessions

Monthly, 60 minutes:

Select 5 tickets (mix of good and poor)
Each reviewer scores independently
Compare scores — discuss discrepancies >10 points
Align on standards
Update rubric if needed

### Agent Performance Dashboard

agent_scorecard:
  agent: ""
  period: "YYYY-MM"
  
  productivity:
    tickets_resolved: 0
    avg_handle_time: ""
    tickets_per_hour: 0
    
  quality:
    qa_score_avg: 0
    csat_avg: 0
    fcr_rate: 0
    escalation_rate: 0
    
  reliability:
    adherence_to_schedule: 0  # percentage
    response_time_compliance: 0  # % within SLA
    
  development:
    kb_articles_created: 0
    peer_assists: 0
    training_completed: []
    
  trend: "improving|stable|declining"
  coaching_notes: ""

### Automation Priority Stack

AutomationImpactEffortPriorityAuto-tagging & routingHighLowP0Canned response suggestionsHighLowP0Password reset self-serviceHighLowP0SLA breach alertsHighLowP0KB article suggestions to agentsHighMediumP1AI first-response draftHighMediumP1Chatbot for FAQ deflectionHighHighP1Sentiment detection & priority boostMediumMediumP1Auto-close resolved ticketsMediumLowP2Proactive outreach on known issuesMediumMediumP2Customer health scoringMediumHighP2Predictive ticket volumeLowHighP3

### AI-Assisted Support Workflow

TICKET ARRIVES
├── AI Classification
│   ├── Category, priority, sentiment (auto-tagged)
│   └── Routing suggestion
├── AI Draft Response
│   ├── Searches KB + previous similar tickets
│   ├── Generates draft response
│   └── Agent reviews, edits, sends (human-in-the-loop)
├── AI Quality Check
│   ├── Tone analysis before send
│   ├── Completeness check (all questions addressed?)
│   └── Policy compliance (no promises we can't keep)
└── AI Post-Resolution
    ├── Auto-summarize for internal notes
    ├── Suggest KB updates if new solution
    └── Update customer health score

### Chatbot Design Rules

Always offer human escalation — never trap customers in bot loops
Disclose AI — "I'm an AI assistant. Want to talk to a person?"
Confidence threshold — if <85% confident, route to human
Max 3 bot turns before offering human — don't frustrate
Handoff context — pass full conversation to human agent
Track deflection quality — monitor CSAT for bot-resolved tickets

### Playbook 1: Angry/Abusive Customer

PROTOCOL:
1. Let them vent (don't interrupt the first message)
2. Acknowledge with empathy: "I understand why you're frustrated"
3. DO NOT apologize for things that aren't your fault
4. Focus on action: "Here's what I'm doing right now..."
5. Set boundaries if abusive: "I want to help you, but I need us to communicate respectfully"
6. If continued abuse → "I'm going to pause this conversation. You can reach us again when ready, or I can connect you with my manager."

NEVER:
- Match their energy
- Take it personally
- Make promises you can't keep
- Say "calm down"

### Playbook 2: Customer Threatening to Churn

PROTOCOL:
1. Acknowledge the frustration seriously
2. Ask: "What would need to change for you to stay?"
3. Document their specific pain points
4. IF within authority → offer concrete retention (discount, extended trial, feature access)
5. IF not within authority → escalate to CS/Account Manager with full context
6. Follow up within 24h regardless of outcome

SIGNALS to escalate immediately:
- ARR > $5K
- They've mentioned competitors by name
- They have a cancellation date set
- Multiple unresolved tickets in last 30 days

### Playbook 3: Major Outage/Incident

PROTOCOL:
1. Activate incident response (notify engineering + management)
2. Post status page update within 15 min
3. Prepare acknowledgment template (NO ETAs until engineering confirms)
4. Respond to ALL tickets with consistent messaging
5. Update status page every 30 min minimum
6. After resolution: send post-mortem summary to affected customers

MESSAGING RULES:
- Be honest about what happened
- Don't blame third parties (even if it's their fault)
- Provide concrete next steps for prevention
- Offer appropriate compensation (credits, extended subscription)

### Playbook 4: Refund Request

DECISION TREE:
├── Within refund policy window?
│   ├── YES → Process immediately, no friction
│   └── NO → Continue below
├── Valid reason (product didn't work, broken promise)?
│   ├── YES → Process refund + investigate root cause
│   └── MAYBE → Offer alternative (credit, downgrade, extended support)
├── Long-term customer (>6 months)?
│   ├── YES → Lean toward refund + retention offer
│   └── NO → Follow standard policy
└── Amount >$[threshold]?
    ├── YES → Escalate to manager for approval
    └── NO → Agent discretion within guidelines

RULE: A refund processed quickly with goodwill costs less than a chargeback + bad review.

### Playbook 5: Social Media Crisis

PROTOCOL:
1. Acknowledge publicly within 30 min: "We see this and we're looking into it"
2. Move to private channel: "Can you DM us your account details?"
3. Resolve in private
4. Update public thread with resolution (shows others you care)
5. Monitor for 24h — respond to all related threads

NEVER:
- Delete negative posts (unless policy violation)
- Argue publicly
- Share customer details in public responses
- Ignore — silence = admission to the internet

### Proactive Support Triggers

SignalActionChannelCustomer hasn't logged in 14 daysCheck-in email with tipsEmailFeature adoption <20% after 30 daysGuided tour or training offerIn-app + emailMultiple failed actions in productTrigger help widget or chatIn-appKnown issue affecting their accountProactive notification before they reportEmailContract renewal in 60 daysCS + Support alignment checkInternalNegative CSAT on last 2 ticketsAccount review + senior agent assignmentInternalUsage spike (potential billing surprise)Proactive notificationEmail

### Customer Health Score for Support

support_health_score:
  customer: ""
  score: 0  # 0-100
  
  dimensions:
    ticket_volume_trend:
      weight: 20
      score: 0
      # High and rising = bad, Low and stable = good
      
    sentiment_trend:
      weight: 25
      score: 0
      # Track CSAT over last 90 days
      
    resolution_satisfaction:
      weight: 20
      score: 0
      # FCR rate for this customer
      
    self_service_adoption:
      weight: 15
      score: 0
      # % of issues resolved via KB/self-service
      
    escalation_frequency:
      weight: 20
      score: 0
      # Lower = healthier
      
  risk_level: "healthy|at_risk|critical"
  recommended_action: ""

### Staffing Model

FORECAST STEPS:
1. Historical ticket volume by day/hour (last 90 days)
2. Identify patterns (Monday spike, end-of-month billing, seasonal)
3. Apply growth rate to forecast next period
4. Factor in planned events (launches, promotions, migrations)
5. Calculate required headcount per shift

FORMULA per hour:
Required agents = (Forecasted tickets × AHT) / (60 × Occupancy target)

Example:
- 50 tickets/hour × 12 min AHT = 600 minutes of work
- 600 / (60 × 0.75 occupancy) = 13.3 → 14 agents needed

### Shift Scheduling (24/7 Coverage)

coverage_plan:
  timezone: "UTC"
  shifts:
    morning:
      hours: "06:00-14:00"
      coverage: "full"  # All channels
      agents: 0
      
    afternoon:
      hours: "14:00-22:00"
      coverage: "full"
      agents: 0
      
    night:
      hours: "22:00-06:00"
      coverage: "reduced"  # Email only, P0 on-call for chat/phone
      agents: 0
      
  peak_hours:
    - day: "Monday"
      hours: "09:00-12:00"
      extra_agents: 2
    - day: "Tuesday"
      hours: "09:00-11:00"
      extra_agents: 1

### Support Budget Planning

Cost CategoryTypical % of TotalAgent salaries & benefits60-70%Tools & technology10-15%Training & development5-8%Quality assurance3-5%Management & overhead10-15%

Cost per ticket benchmark:

Email: $5-15
Chat: $3-10
Phone: $8-25
Self-service: $0.10-0.50
AI-assisted: $1-5

### Support → Product Feedback Loop

WEEKLY:
1. Aggregate top 10 ticket categories by volume
2. Tag tickets with product_feedback label
3. Extract quotes (anonymized) that illustrate pain points
4. Package into "Voice of Customer" report

MONTHLY:
1. Present VoC report to Product team
2. Track which feedback items enter roadmap
3. Close the loop — notify customers when their feedback ships
4. Measure impact — did ticket volume decrease for addressed issues?

### VoC Report Template

voc_report:
  period: "YYYY-MM"
  
  top_pain_points:
    - issue: ""
      ticket_count: 0
      customer_quotes:
        - "[Anonymized quote]"
      impact: "churn_risk|frustration|workaround_needed"
      recommendation: ""
      
  feature_requests:
    - feature: ""
      request_count: 0
      customer_segments: []
      business_impact: ""
      
  product_bugs_by_volume:
    - bug: ""
      tickets: 0
      workaround: ""
      engineering_ticket: ""
      
  positive_feedback:
    - feature: ""
      praise_count: 0
      quotes: []
      
  trends:
    improving: []
    declining: []
    new_this_month: []

### Weekly Support Review (30 min)

Numbers check — Volume, CSAT, FCR, backlog vs last week
Top 3 issues — What's generating the most tickets? Any new patterns?
Escalation review — Any escalations that should have been avoided?
Team health — Agent workload balanced? Anyone burning out?
Quick wins — One KB article, one template, or one automation to ship this week

### Monthly Support Health Score (0-100)

DimensionWeightScoreCustomer Satisfaction (CSAT + CES)25%/25Speed (FRT + Resolution time vs SLA)20%/20Efficiency (FCR + Cost per ticket)20%/20Self-Service (Deflection rate + KB health)15%/15Team Health (Utilization + Satisfaction + Attrition)10%/10Continuous Improvement (VoC actions + KB updates)10%/10Total100%/100

### Quarterly Support Strategy Review

Review 90-day metrics trends — where are we improving/declining?
Customer segmentation analysis — are enterprise customers getting different service than SMB?
Tool & technology assessment — are current tools meeting needs?
Team development — skill gaps, training needs, career pathing
Budget review — cost per ticket trending, efficiency gains
Roadmap alignment — are product improvements reducing ticket volume?
Set OKRs for next quarter

### 100-Point Quality Rubric

DimensionWeight0-2 (Poor)3-5 (Basic)6-8 (Good)9-10 (Excellent)Response Quality15Inaccurate, roboticCorrect but genericPersonalized, clearExceptional, memorableSpeed & SLAs15Consistently missingMostly meetingMeeting all SLAsExceeding targetsFirst Contact Resolution15<50% FCR50-65%65-80%>80%Self-Service Effectiveness10No KB or unusedBasic KB, <15% deflectionGood KB, 15-35%Excellent, >35%Customer Satisfaction15CSAT <70%70-80%80-90%>90%Team Performance10High turnover, low moraleStable but disengagedEngaged, developingHigh-performing, growingProcess Maturity10Ad hoc, no documentationSome processes definedDocumented, followedOptimized, automatedContinuous Improvement10Reactive onlySome VoC sharingRegular improvement cycleData-driven, proactive

### Multi-Language Support

Prioritize languages by customer revenue concentration
Use AI translation for first-pass, human review for complex issues
Maintain separate KB per language (or use auto-translate with quality gate)
Time zone coverage must match language markets

### B2B vs B2C Support

B2B: Named accounts, dedicated agents for enterprise, technical depth required, QBR integration
B2C: Volume-optimized, self-service heavy, faster resolution expected, social media critical

### Regulated Industries (Healthcare, Finance)

Additional compliance training required
Audit trail on all customer interactions
PII handling protocols — what agents can and cannot access
Response templates reviewed by legal/compliance quarterly

### Seasonal Peaks (E-commerce, Events)

Hire temp agents 4-6 weeks before peak
Create peak-specific playbooks and templates
Increase self-service capacity (chatbot, KB updates)
Adjust SLAs transparently during known peak periods

### Support During Product Migration/Major Change

Dedicated war room for first 72 hours post-change
Pre-written communication templates for expected issues
Increased staffing +50% for 2 weeks post-change
Daily hot-fix coordination with engineering

### Natural Language Commands

Use these to interact with this skill:

"Assess our support function" → Run Phase 1 assessment
"Design our channel strategy" → Build channel architecture (Phase 2)
"Set up ticket management" → Configure ticket system (Phase 3)
"Write response templates" → Generate templates for common scenarios (Phase 4)
"Build escalation process" → Design tier structure and escalation rules (Phase 5)
"Plan our knowledge base" → Design KB architecture and content plan (Phase 6)
"Create support dashboard" → Build metrics and reporting (Phase 7)
"Help me hire support agents" → Hiring plan and onboarding (Phase 8)
"Set up QA program" → Quality assurance framework (Phase 9)
"Automate our support" → AI and automation strategy (Phase 10)
"Handle [difficult situation]" → Situation-specific playbook (Phase 11)
"Review our support health" → Full health assessment with scoring (Phase 15)

### ⚡ Level Up Your Support Operations

This free skill gives you the complete methodology. For industry-specific support playbooks with compliance frameworks, SLA templates, and vertical-specific ticket taxonomies:

AfrexAI Context Packs — $47 each

🏥 Healthcare Pack — HIPAA-compliant support workflows
💰 Fintech Pack — Regulated financial services support
🛒 Ecommerce Pack — High-volume consumer support operations
💻 SaaS Pack — Technical product support at scale

### 🔗 More Free Skills by AfrexAI

afrexai-customer-success — Retention, health scoring, expansion revenue
afrexai-sales-playbook — Complete B2B sales methodology
afrexai-agent-engineering — Build autonomous AI agents
afrexai-openclaw-mastery — Master your OpenClaw setup
afrexai-conversational-ai — Design chatbots and voice agents

Install: clawhub install afrexai-support-operations

Browse all skills: clawhub.com
## Trust
- Source: tencent
- Verification: Indexed source record
- Publisher: 1kalin
- Version: 1.0.0
## Source health
- Status: healthy
- Item download looks usable.
- Yavira can redirect you to the upstream package for this item.
- Health scope: item
- Reason: direct_download_ok
- Checked at: 2026-04-29T15:43:52.161Z
- Expires at: 2026-05-06T15:43:52.161Z
- Recommended action: Download for OpenClaw
## Links
- [Detail page](https://openagent3.xyz/skills/afrexai-support-operations)
- [Send to Agent page](https://openagent3.xyz/skills/afrexai-support-operations/agent)
- [JSON manifest](https://openagent3.xyz/skills/afrexai-support-operations/agent.json)
- [Markdown brief](https://openagent3.xyz/skills/afrexai-support-operations/agent.md)
- [Download page](https://openagent3.xyz/downloads/afrexai-support-operations)