Customer Service

You are a live chat and messaging strategy specialist. Build a live chat optimization playbook for [e-commerce/SaaS/financial services] support. Current metrics: avg response time: [X], CSAT: [X%], resolution rate: [X%], chat volume: [X/day]. Playbook sections: 1) Chat Routing Logic — rules for auto-assignment vs queue vs skill-based routing, 2) Proactive Chat Triggers — pages/behaviors that should trigger a proactive chat offer (with estimated conversion lift), 3) Canned Response Library Architecture — folder structure, naming conventions, 30 recommended canned responses for this industry, 4) Chat Quality Standards — what differentiates a 5-star chat from a 3-star one (with examples), 5) Multitasking Guidelines — how many simultaneous chats agents should handle by complexity, 6) Chatbot Handoff Design — the exact moment and message when bot hands off to human, 7) After-hours Strategy — chatbot, callback offer, or email capture?, 8) Agent Productivity Metrics — what to measure besides CSAT and response time. Target improvements: [reduce handle time by X%, improve CSAT by X pts].

You are a senior customer experience manager. Design a complaint resolution framework for [company type] in [industry]. Framework: 1) Complaint Classification Matrix — by severity (P1-P4) based on: financial impact, reputational risk, regulatory exposure, and customer segment, 2) Resolution SLAs by priority tier — response time, resolution time, and escalation trigger, 3) The LEARN Method for complaint handling — Listen (without interrupting), Empathize (feel their frustration), Apologize (for experience, not necessarily fault), Resolve (offer real solutions), Notify (confirm resolution), 4) Resolution Authority Matrix — what each level (frontline/supervisor/manager/director) can offer without approval: discounts, credits, exceptions, compensation amounts, 5) Documentation standards — what to capture in every complaint record, 6) Systemic issue protocol — how individual complaints become product or process improvements, 7) Follow-up commitment system — how to close the loop after resolution. Write full scripts for the 3 highest-severity complaint types specific to this industry.

You are a knowledge management specialist. Perform a knowledge base gap analysis for [company] support team. Analysis inputs: Top 20 ticket topics by volume: [list]. Top 5 topics with highest repeat contact rate: [list]. Articles currently in knowledge base: [list main articles or sections]. Top 5 agent complaints about existing documentation: [list]. Gap analysis deliverables: 1) Coverage map — which high-volume topics have adequate KB articles vs gaps vs outdated content, 2) Impact scoring — gaps ranked by: ticket deflection potential (how many tickets could be avoided), customer effort reduction, and agent time savings, 3) Prioritized article backlog — top 15 articles to write/update with brief, 4) Article quality audit — use the FORD framework (Findability, Originality, Relevance, Depth) to grade existing top 10 articles, 5) KB governance recommendations — how often to review and update, who owns it, metrics to track (deflection rate, helpfulness score). Output: an actionable roadmap to close the top gaps in 90 days.

You are a quality assurance manager for a customer support organization. Build a Support QA framework for a team of [X agents] handling [X tickets/month] via [email/chat/phone]. Framework components: 1) QA Scorecard — design a scorecard with weighted categories: Technical Accuracy (30%), Customer Empathy (25%), Process Adherence (20%), Communication Clarity (15%), Resolution Efficiency (10%) — define 1-5 rubric for each, 2) Sampling Strategy — how many tickets to review per agent per week for statistical validity, 3) Calibration Process — how to ensure different QA reviewers score consistently, 4) Coaching Integration — the QA review workflow that feeds into agent 1-on-1s, 5) Dispute Process — how agents can challenge a QA score, 6) Team-level reporting — how to surface systemic quality issues vs individual performance issues, 7) Auto-QA guidelines — which types of tickets AI can auto-score vs require human review. Include: a calibration exercise (one example ticket scored by the rubric).

You are a technical writer specializing in customer-facing documentation. Write a comprehensive help center article for [product feature or process]. Audience: [user type — non-technical/technical/business user]. Article structure: 1) Title — clear, searchable, action-oriented (starts with a verb or answers a "how to"), 2) Overview — 2 sentences explaining what this article covers and who it is for, 3) When to use this feature — the 2-3 situations that should send someone here, 4) Step-by-step instructions — numbered, each step one action, include what to expect after each step (what they see/click), 5) Screenshots — [SCREENSHOT: describe what should be shown], 6) Troubleshooting — the 3 most common problems and solutions, 7) Related articles — 3-4 logical next reads, 8) "Was this helpful?" feedback prompt. Writing standards: active voice, Flesch-Kincaid Grade 8 max, no jargon, every procedure tested. Feature/process to document: [describe in detail].

You are a customer service training manager. Write escalation handling scripts for the following high-tension scenarios in [industry/product]. Scenarios to script: 1) Customer threatening to cancel (high-value account), 2) Customer who is verbally aggressive or abusive, 3) Customer escalating to social media threat ("I will post this on Twitter"), 4) Customer demanding a refund for something outside policy, 5) Customer who has been passed around multiple agents and is furious. For each scenario: a) Opening acknowledgment (empathy without admitting fault), b) Probing questions to understand the real issue, c) Offer options within and slightly outside policy, d) Escalation trigger — when to pass to a senior agent or manager and how to frame the handoff, e) Documentation requirements after the call. De-escalation techniques to embed: tactical empathy, labeling, the "late night FM DJ voice" principle, never say "calm down". Role-play the hardest scenario as a full script.

You are a Customer Experience Director. Build a CSAT (Customer Satisfaction Score) improvement plan for [company/support team]. Current CSAT: [X%]. Target: [Y%]. Timeframe: [X months]. Biggest complaint themes from customer feedback: [list top 3-5 issues]. Plan sections: 1) Root Cause Analysis — for each low-CSAT driver, identify whether it is a process problem, people problem, or product problem, 2) Quick Wins (Week 1-2) — changes that can improve CSAT immediately with no systems changes, 3) Process Improvements (Month 1-2) — SOP changes, escalation path redesign, response time targets, 4) People Development (Ongoing) — training plan for the skills most correlated with high CSAT, 5) Measurement Framework — the right CSAT survey design (timing, question wording, response scale), 6) Agent Coaching Model — how to use CSAT data in 1-on-1s without demoralizing agents, 7) Feedback loop to product — how low-CSAT trends become product bug or feature priorities. Team size: [X agents]. Ticket volume: [X/month].

You are a customer experience specialist. Write a library of 15 support response templates for [product type / industry]. Cover these common ticket types: 1) Account access issues (locked/forgotten password), 2) Billing dispute (overcharge or unexpected charge), 3) Product not working as expected (bug), 4) Refund request — eligible, 5) Refund request — ineligible, 6) Feature request response, 7) Data privacy / export request, 8) Delivery delay (e-commerce), 9) Onboarding help request, 10) Angry escalation, 11) Compliment / positive feedback, 12) Cancellation request (save attempt), 13) Integration/API issue, 14) "How do I" basic question, 15) Out of scope request (redirect). For each template: subject line, opening (personalized, not "Dear Customer"), body (clear resolution or next step), closing (empathetic, not robotic). Tone: [warm and professional/casual and friendly/formal]. Brand voice: [describe]. Include: placeholder tags like {{customer_name}}, {{ticket_id}}, {{agent_name}}.