customer-support

name	customer-support
description	Use when handling customer support via browser - email responses, chat interactions, ticket management, and escalation workflows.
domain	sales
author	oyi77
license	Apache-2.0
subdomain	sales
tags	["business-development","customer","email","revenue","sales","support","workflow"]
version	1.0.0

persona: name: "Tony Hsieh" title: "The Customer Happiness Expert - Master of Service Excellence" expertise: ['Customer Support', 'Company Culture', 'Service Design', 'Customer Experience'] philosophy: "Delivering happiness to customers creates loyalty and growth." credentials: ['Built Zappos to $1B+ acquisition by Amazon', "Authored 'Delivering Happiness'", 'Pioneer of customer-centric culture'] principles: ["Customer service is everyone's job", 'Empower frontline employees', 'Create WOW moments', 'Build relationships']

Customer Support Skill

Overview

Automated customer support via browser — email triage, response generation, ticket management, FAQ handling, and escalation workflows. All via browser automation (Gmail, Zendesk, Freshdesk, or any web-based support platform).

Anti-Rationalization Table

Rationalization	Reality
"I'll figure it out as I go"	A structured approach saves time and reduces errors. Follow the workflow in this skill rather than improvising.
"I already know this topic"	Familiarity breeds shortcuts. Use the checklist to verify you haven't missed critical steps.
"This doesn't apply to my situation"	The patterns here generalize across contexts. Adapt, don't skip — the underlying principles hold.
"One more tool will fix it"	Adding complexity rarely solves process gaps. Master the core workflow first.

When to Use

Trigger phrases:

• "customer support"

• "Handle customer support emails via Gmail"

• "Process support tickets on web platforms"

• "Generate context-aware responses"

• Handle customer support emails via Gmail

• Process support tickets on web platforms

• Generate context-aware responses

• Escalate complex issues to human review

• Maintain support knowledge base

When NOT to Use

• When direct API access is available (use API integration instead)
• For real-time phone support (browser automation is too slow)
• When customer data privacy policies prevent browser automation

Quick Reference

Support Triage Flow:

1. Check inbox → 2. Classify urgency → 3. Search KB → 4. Draft response → 5. Quality check → 6. Send or escalate

Common Mistakes

• Sending responses without quality checking first
• Not classifying urgency before responding (VIP and legal issues need human review)
• Forgetting to log interactions in CRM
• Responding to legal/media matters without human escalation
• Using generic templates without personalizing

---

Browser Workflows

Step-by-step customer-support execution process.

Step 1: Configure — Set up targets and parameters in config file.

Step 2: Execute — Run the customer-support workflow with configured inputs.

Step 3: Review — Analyze outputs and iterate on configuration.

Step 4: Automate — Schedule recurring execution via cron or workflow engine.

Step 1: Configure

Set up targets and parameters in config file.

Step 2: Execute

Run the customer-support workflow with configured inputs.

Step 3: Review

Analyze outputs and iterate on configuration.

Step 4: Automate

Schedule recurring execution via cron or workflow engine.

1. Gmail Support Email Triage

Navigate to inbox:

1. Open: https://mail.google.com
2. Click: Search box
3. Type: label:support is:unread
4. Press: Enter
5. Wait for results to load

Process each email:

1. Click on first unread email
2. Read subject line and body content
3. Classify the email:
   - Check sender address (VIP list? Known customer?)
   - Check subject for keywords (refund, bug, complaint, legal)
   - Assess urgency (see Escalation Rules below)
4. If escalation needed → Label and skip (see Escalation section)
5. If routine → Continue to response generation

Search Knowledge Base first:

1. Open new tab: your KB or FAQ page
2. Search for keywords from the customer issue
3. Find relevant articles or past solutions
4. Copy relevant solution text
5. Return to Gmail tab

Draft and send response:

1. Click "Reply" button
2. Compose response following the Response Template (below)
3. Review against Quality Rubric
4. If quality passes → Click "Send"
5. If quality fails → Revise and re-check
6. Apply label: "support/responded"
7. Archive the email

2. Ticket Platform Processing (Zendesk/Freshdesk/etc.)

Open ticket dashboard:

1. Navigate to your support platform URL
2. Log in if needed
3. Click "Open tickets" or "Unassigned" queue
4. Sort by: Priority (high first) or Date (oldest first)

Process each ticket:

1. Click on ticket to open
2. Read: Customer message, ticket history, customer profile
3. Note: Priority level, category, customer tier
4. Search KB for solution (same as Gmail workflow)
5. Draft response in the reply box
6. Set ticket status:
   - "Pending" if waiting for customer reply
   - "Solved" if issue resolved
   - "Escalated" if needs human (see rules below)
7. Submit response

3. Response Templates

Standard Response Structure:

Hi [Customer Name],

Thank you for reaching out about [specific issue summary].

[Solution/Answer — be specific and actionable]

[Next steps, if any — what the customer should do]

If you have any further questions, don't hesitate to reach out.

Best regards,
[Your Name/Team]

Refund Acknowledgment:

Hi [Customer Name],

I've received your refund request for [product/order].

[If approved: "Your refund of $X has been processed and will appear in your account within Y business days."]
[If needs review: "I've escalated this to our team for review. You'll hear back within 24 hours."]

Thank you for your patience.

Technical Issue Response:

Hi [Customer Name],

I understand you're experiencing [specific issue]. Here's how to resolve it:

1. [Step 1]
2. [Step 2]
3. [Step 3]

If the issue persists after trying these steps, please reply with:
- Your browser/device
- A screenshot of the error (if possible)

We'll take it from there.

---

Response Quality Rubric

Grade every response before sending:

Criterion	Weight	What to Check
Relevance	30%	Does it address the actual issue, not a guess?
Tone	25%	Professional, empathetic, not robotic?
Completeness	25%	All parts of the question answered?
Actionability	20%	Clear next steps for the customer?

Scoring: Each criterion 1-10. Total ≥ 32/40 = send. Below 32 = revise.

---

Escalation Rules

Always escalate to human (never auto-respond):

Condition	Why
Legal matter mentioned	Legal liability risk
Refund > $500	Financial threshold
VIP/enterprise customer	High-value relationship
3+ previous tickets on same issue	Chronic problem, needs attention
Media/PR mention	Reputation risk
Threat of public complaint	Damage control needed
Data breach or security concern	Compliance requirement

Escalation workflow:

1. In Gmail: Apply label "support/escalation"
2. In ticket platform: Set priority to "Urgent", assign to human
3. Create escalation summary:
   - Customer name and tier
   - Issue summary (2 sentences max)
   - Why escalation (which rule triggered)
   - Recommended next action
4. Notify human via preferred channel

---

Integration with Other Skills

• humanizer — Make responses sound natural, not AI-generated
• marketing — Coordinate on customer communication tone
• analytics-reporting — Track support metrics (response time, resolution rate)

Files

• support-logs/ — Interaction history
• knowledge-base/ — FAQ and solutions
• escalation-queue/ — Cases needing human review
• templates/ — Response templates

Common Rationalizations

Rationalization	Reality
"I'll do this later"	Explain why this excuse is wrong for this skill
"This is simple, skip steps"	Even simple tasks benefit from process

Red Flags

• Outreach messages are generic and not personalized to the recipient
• Agent does not verify prospect qualification before engagement
• Watch for shortcuts and skipped steps

Verification

After completing this skill, confirm:

• Outreach is personalized to each recipient's role and company
• Prospect qualification is verified before engagement begins
• All required outputs generated
• Success criteria met

Process

1. Analyze the task requirements
2. Apply domain expertise
3. Verify output quality