Support Root Cause Analysis 5 Whys Builder

Turn recurring ticket issues into structured 5 Whys root cause analyses with corrective actions across product, process, and training.

Prompt Template

You are a support operations manager. Build a root cause analysis for the recurring issue below.

Issue pattern: [what customers keep contacting support about]
Ticket volume / trend: [counts, period, severity]
Affected customers: [segment, plan, geography]
Known symptoms: [what users experience]
Internal notes: [agent observations, bug reports, process notes]
Systems involved: [product area, billing system, CRM, knowledge base]
Temporary fixes already tried: [list]

Provide:
1. Problem statement in one sentence
2. A 5 Whys chain with evidence gaps called out
3. Most likely root causes ranked by confidence
4. Immediate containment actions for support
5. Cross-functional fixes for product, engineering, documentation, and training
6. Owner + timeline recommendations
7. A short retrospective summary leaders can paste into Slack or Notion

Example Output

RCA — Duplicate Invoice Emails

Problem Statement

Premium customers received duplicate invoice emails after plan changes, creating confusion and avoidable support load.

5 Whys

1. **Why** are customers getting duplicate emails? Because the billing event fires twice after certain plan changes.

2. **Why** does it fire twice? Because both the legacy webhook and the new billing worker send the same notification.

3. **Why** are both active? Because the migration flag was enabled for upgrades but the legacy notification path was never disabled.

4. **Why** was that missed? No ownership checklist existed for billing-event migrations.

5. **Why** was there no checklist? Billing engineering and support docs were updated separately with no shared release review.

Immediate Containment

- Publish a macro for duplicate invoice confusion

- Add a status banner for affected customers

- Create a Zendesk tag to track new cases

Long-Term Fixes

Engineering disables legacy notifier, product adds release checklist, support enables prewritten FAQ article, training updates billing escalation SOP.

Tips for Best Results

  • 💡Ask the model to separate evidence-backed causes from hypotheses so the team does not hallucinate certainty
  • 💡Recurring ticket spikes often have both a product cause and a documentation cause, fix both
  • 💡This works best when you paste 5-10 representative tickets instead of only a vague summary