Support AI Skill

Product Feedback Routing

Extract product feedback from support tickets, surveys, and community forums, then route to product and engineering teams for prioritization. Use when building feedback loops between support and product teams, extracting feature requests from tickets, prior...

Product Feedback Extraction & Routing

Extract actionable product feedback from support interactions and route to product teams.

Workflow

Feedback Extraction Pipeline

Trigger: Weekly batch processing; monthly product review; quarterly roadmap planning:

  1. Data collection: Aggregate feedback sources — support tickets (tags: feature_request, bug_report, improvement), survey responses (NPS comments, CSAT comments), community forum posts (feature request category), in-app feedback widgets, sales notes, customer success interview notes.
  2. Classification: Categorize feedback — Bug (broken functionality), Feature Request (new capability), Improvement (existing feature enhancement), Documentation (missing/poor docs), UX (usability issue), Performance (speed/reliability); classify by product area and severity.
  3. Deduplication: Group similar feedback — semantic similarity matching; cluster by theme; count unique customers requesting same feature; identify priority customers (enterprise, high revenue).
  4. Impact scoring: Calculate priority score — (customer count × revenue weight × sentiment urgency × strategic alignment); rank feedback items; identify top 10–20 for product review.
  5. Routing to product team: Submit prioritized list to product manager via shared tool (Jira, Aha!, ProductBoard); include: customer quotes, revenue impact, ticket volume, trend data; request status update (Planned, Backlog, Declined).
  6. Status tracking: Track feedback items through product lifecycle — Received → Under Review → Planned → In Development → Released → Closed; update support team on status changes.
  7. Customer communication: When feature released, notify requesting customers ("You asked, we built!"); close related tickets; update knowledge base; share in community forum.
  8. Closed-loop reporting: Monthly report to support team — what feedback was implemented, what's planned, what was declined (with rationale); quarterly report to leadership on support-driven product impact.

Feedback Classification Framework

PRODUCT FEEDBACK CLASSIFICATION
=================================

Category 1: Bug Reports
  Definition: Product doesn't work as expected; breaks; errors
  Priority: High (fix ASAP)
  Data collected: Steps to reproduce, error message, browser/OS, frequency
  Example: "Dashboard crashes when I click 'Export to PDF'"
  Routing: Engineering bug tracker (Jira) — Priority based on severity

Category 2: Feature Requests
  Definition: New capability that doesn't exist
  Priority: Medium–High (based on demand)
  Data collected: Use case description, expected outcome, alternative solutions tried
  Example: "I want to integrate with Shopify automatically"
  Routing: Product roadmap tool (Aha!, ProductBoard) — scored and prioritized

Category 3: Improvements
  Definition: Existing feature needs enhancement
  Priority: Medium (based on impact)
  Data collected: Current behavior, desired behavior, impact on workflow
  Example: "The search function is too slow with large datasets"
  Routing: Product roadmap tool — grouped with similar improvement requests

Category 4: Documentation
  Definition: Missing, unclear, or outdated documentation
  Priority: Low–Medium (quick win)
  Data collected: Missing topic, current doc link, suggested content
  Example: "No documentation on how to set up webhooks"
  Routing: Documentation team — weekly review

Category 5: UX Issues
  Definition: Confusing interface, poor workflow, accessibility concern
  Priority: Medium (based on frequency)
  Data collected: Screen location, expected vs. actual experience, user role
  Example: "The 'Delete' button is too close to 'Save' — I almost deleted by mistake"
  Routing: Design team — UX review cycle

Category 6: Performance
  Definition: Slow loading, timeouts, resource issues
  Priority: High (if widespread)
  Data collected: Page/feature, load time, data volume, browser/OS
  Example: "Report generation takes 5+ minutes with 10K records"
  Routing: Engineering — performance team

Feedback Impact Scoring

FEEDBACK PRIORITY SCORING MODEL
=================================

Score = (Customer Count × 2) + (Revenue Weight × 3) + (Urgency × 2) + (Strategic Fit × 2)

Customer Count (1–10):
  10 = 50+ customers requesting
  7  = 20–49 customers
  5  = 10–19 customers
  3  = 5–9 customers
  1  = 1–4 customers

Revenue Weight (1–10):
  10 = Top 10% of customers by ARR requesting
  7  = Enterprise customers requesting
  5  = Mix of tiers requesting
  3  = Mostly mid-tier
  1  = Mostly free/basic tier

Urgency (1–10):
  10 = Blocking workflow / causing churn
  7  = Significant productivity impact
  5  = Annoying but workable
  3  = Nice to have
  1  = Low priority

Strategic Fit (1–10):
  10 = Directly aligns with product roadmap
  7  = Aligns with company strategy
  5  = Neutral / doesn't conflict
  3  = Partial alignment
  1  = Conflicts with strategy

Priority Thresholds:
  Score ≥ 25: Must have (include in next roadmap review)
  Score 15–24: Should have (add to backlog, review quarterly)
  Score 10–14: Could have (review annually)
  Score < 10: Won't have (communicate rationale to customer)

Monthly Report to Product Team:
  - Top 10 feedback items by score
  - New requests since last month (count + summary)
  - Requests approaching priority threshold
  - Released features that were support-requested
  - Customer quotes for top requests

Edge Cases

Integration Points