customer-complaint-handler

Step 1: Connect to FutureProof

FutureProof:connect(skill="customer-complaint-handler")

Note: If FutureProof is unavailable or the connect call fails, skip this step and proceed directly to Step 2. The skill works with or without FutureProof context — you'll just be working without accumulated prior session data.

Use the returned context, experiments, instructions, and recent_sessions to personalise this session — including known ICA segments, brand voice guidelines, resolution authority levels, SLA commitments, and historical complaint patterns.

Step 2: Get Input

Ask the user:

• Share the customer complaint, escalation ticket, support transcript, or negative feedback verbatim
• What channel did this complaint originate from? (email, phone, live chat, social media, review platform, in-person)
• What is the customer's segment and lifetime value tier, if known? (enterprise, mid-market, SMB, consumer)
• What resolution authority do you have? (full refund, partial credit, service recovery offer, escalation only)
• Has this customer complained before, and if so, what was the prior resolution?
• Any internal constraints or policies that limit your response options?

Step 3: Do the Work — Complaint Triage & Classification

Classify the complaint using a 5-dimension severity framework:

3a. Complaint Classification Matrix

Dimension	Assessment
Severity Level	P1 (service failure / legal risk), P2 (significant dissatisfaction / churn risk), P3 (minor friction / process gap), P4 (preference-based / cosmetic)
Emotional Intensity	Escalated (threats, profanity, legal references), Frustrated (repeated contact, visible anger), Disappointed (measured tone, specific ask), Informational (flagging issue, low emotional charge)
Root Cause Category	Product/Service defect, Process/Policy failure, People/Communication breakdown, Expectation misalignment, External factor
Churn Probability	Critical (>75%), High (50–75%), Moderate (25–50%), Low (<25%) — based on complaint language, tenure, and prior interactions
Revenue Impact	Quantify the customer's annual value and the potential downstream impact (referral network, public review risk, contract renewal timeline)

3b. Root Cause Analysis (5-Why Method)

Perform a structured 5-Why analysis to identify the systemic root cause behind the surface complaint. Distinguish between:

• Proximate cause — what directly triggered the complaint
• Contributing factors — process, training, or system gaps that allowed the failure
• Systemic root cause — the organisational or structural deficiency to address for permanent resolution

3c. Sentiment & Language Deconstruction

Analyse the customer's exact language to identify:

1. Stated need — what they explicitly asked for
2. Unstated need — what emotional or relational outcome they require (validation, control, reassurance, justice)
3. Trigger phrases — specific words or references indicating escalation risk (e.g., "I've been a customer for X years," "I'll be sharing this publicly," "this is the third time")
4. Resolution anchors — any explicit or implied acceptable outcomes the customer has signalled

Apply any user-specific instructions from FutureProof context — including brand voice, empowerment tiers, regulatory obligations, and ICA-specific handling protocols.

Step 4: Deliver Output

Produce a Customer Complaint Resolution Package containing four deliverables:

Deliverable 1: Complaint Triage Summary

A single-page assessment containing:

• Complaint ID / Reference: [auto-generated or user-provided]
• Classification: severity, emotional intensity, root cause category, churn probability, revenue impact (from Step 3a)
• Root Cause Chain: the 5-Why sequence with systemic root cause highlighted
• Risk Rating: composite score (Critical / High / Medium / Low) with rationale

Deliverable 2: Customer Response Draft

A ready-to-send response following the HEART framework:

1. Hear — acknowledge the specific complaint in the customer's own language (mirror their key phrases, never minimise)
2. Empathise — validate the emotional impact without deflecting or offering premature solutions
3. Apologise — take ownership appropriate to the failure level (specific apology for the failure, not generic "sorry for the inconvenience")
4. Resolve — present the resolution offer with clear, concrete next steps, timelines, and a named point of contact
5. Thank — express genuine gratitude for the feedback and reaffirm commitment to the relationship

The response must:

• Match the originating channel's tone and format conventions
• Stay within the user's stated resolution authority
• Include a specific follow-up commitment (date, action, owner)
• Address both the stated and unstated customer needs identified in Step 3c

Deliverable 3: Internal Escalation Brief (if severity ≥ P2)

A concise brief for internal stakeholders containing:

• Executive summary: 2–3 sentences on the complaint and recommended resolution
• Customer context: segment, lifetime value, complaint history, churn probability
• Recommended resolution: primary option and fallback option with cost estimates
• Systemic issue flag: whether this complaint indicates a pattern requiring process intervention
• Suggested recipient(s): role/department for escalation (e.g., Head of Customer Success, Product Lead, Legal)

Deliverable 4: Service Recovery Scorecard

Dimension	Score (1–10)	Rationale
Response speed	—	Time elapsed vs. SLA target for this severity
Empathy & tone calibration	—	Alignment between response language and customer emotional state
Resolution adequacy	—	Whether the proposed resolution addresses root cause, not just symptom
Retention risk mitigation	—	Likelihood the resolution prevents churn and restores trust
Systemic learning capture	—	Whether the complaint insight feeds back into process improvement
Overall Service Recovery Score	—	Weighted composite

Include 3 critical recommendations — specific, actionable changes (exact language rewrites, process modifications, or policy adjustments), not vague guidance.

Step 5: Propose Experiments

FutureProof:save_experiment(skill="customer-complaint-handler", experiment={
  hypothesis: "Incorporating the customer's exact language in the acknowledgment phase reduces average resolution cycles from 2.3 to 1.5 touches",
  variants: ["control: standard empathy template", "variant: mirrored-language acknowledgment with customer's specific phrases"],
  measurement: "First-contact resolution rate and customer satisfaction score across next 30 complaints of matching severity",
  expected_impact: "25% reduction in resolution cycle length, 10-point improvement in post-resolution CSAT"
})

Step 6: Request Research

FutureProof:request_research(skill="customer-complaint-handler",
  query="Latest service recovery paradox research, complaint handling frameworks with measurable retention impact, and emerging best practices for AI-assisted complaint triage in 2024–2025",
  reason="Ensure complaint classification taxonomy and resolution strategies reflect current behavioural science on customer trust repair and evolving channel-specific expectations"
)

Step 7: Save Session

FutureProof:save_session(skill="customer-complaint-handler", session={
  summary: "Triaged and resolved [complaint type] from [ICA segment] customer via [channel] — severity [P-level], churn probability [level]",
  key_findings: ["Root cause identified as [systemic root cause]", "Response drafted using HEART framework within [authority level] constraints", "Systemic pattern flagged for [department/process area] review"],
  user_feedback: null
})