# Ticket Taxonomy Classifier

> Stop manually tagging tickets. This agent reads every ticket, classifies it by topic, product area, urgency, and sentiment, then exports a structured view that surfaces trends your team would otherwise miss.

Source: https://cotera.co/solutions/support/ticket-taxonomy-classifier

---

- **Team:** Support
- **Tools:** Zendesk, Google Sheets
- **Difficulty:** medium
- **Setup time:** 10 min
- **Time saved:** 2 hrs/week

## How it works
1. **Reads incoming tickets** — Pulls tickets from Zendesk with the full message, subject, and any existing tags.
2. **Classifies each dimension** — Assigns topic, product area, urgency, and sentiment based on the ticket content — not just keywords.
3. **Finds taxonomy gaps** — Flags tickets that don't fit your existing categories and suggests new ones based on patterns.
4. **Exports trends** — Builds a Google Sheets report with category breakdowns, trend comparisons, and actionable insights for product and support teams.

## Example requests
- Classify the last 200 tickets by topic, urgency, and sentiment
- Which product area is generating the most frustrated tickets this month?
- Find tickets that don't fit our current taxonomy and suggest new categories

## Prompt

```markdown
## Task

Classify incoming support tickets into a structured taxonomy. For each ticket, assign topic, product area, urgency level, sentiment, and any custom tags. Build a categorized view that surfaces trends and helps prioritize product and support improvements.

## Input

The user provides:
1. A batch of tickets to classify (from Zendesk, Intercom, or a CSV)
2. Their existing taxonomy or category list (or the agent builds one from the data)
3. Optionally, specific tags or dimensions they want to track

**Example:** "Classify the last 200 tickets. Tag each one with: product area (billing, onboarding, integrations, bugs), urgency (low/medium/high/critical), and sentiment (positive/neutral/negative/frustrated)."

## Context

### Classification Dimensions

**Topic (auto-detected):**
- Account & billing
- Onboarding & setup
- Integrations & API
- Bugs & errors
- Feature requests
- How-to questions
- Data & reporting

**Product Area (mapped from topic):**
- Billing system
- Dashboard
- API / webhooks
- Mobile app
- Integrations (specific tool)
- Admin / settings

**Urgency:**
- Critical: Service down, data loss, security issue
- High: Broken workflow, blocking customer
- Medium: Degraded experience, workaround exists
- Low: Question, feature request, cosmetic issue

**Sentiment:**
- Positive: Compliment, satisfied
- Neutral: Factual question, no emotion
- Negative: Complaint, disappointment
- Frustrated: Multiple follow-ups, escalation language

### Classification Process

1. Pull tickets from the specified source
2. Read the initial message and any follow-up context
3. Assign each classification dimension
4. Flag tickets that don't fit existing categories (taxonomy gaps)
5. Aggregate counts per category to surface trends
6. Export the classified data to Google Sheets

## Output

**Classification Summary**

**Tickets Classified:** [count]
**Categories Used:** [count]
**Taxonomy Gaps Found:** [count]

---

**By Topic:**

| Topic | Count | % | Trend vs Last Period |
|-------|-------|---|---------------------|
| Bugs & errors | [n] | [%] | [up/down/flat] |
| Onboarding | [n] | [%] | [up/down/flat] |

---

**By Urgency:**

| Urgency | Count | % |
|---------|-------|---|
| Critical | [n] | [%] |
| High | [n] | [%] |

---

**By Sentiment:**

| Sentiment | Count | % |
|-----------|-------|---|
| Frustrated | [n] | [%] |
| Negative | [n] | [%] |

---

**Taxonomy Gaps:**
- [X] tickets didn't fit existing categories. Suggested new category: [name]
- Common phrases in uncategorized tickets: [phrases]
```

## Related reading
- [Best AI Customer Support Tools](https://cotera.co/articles/best-ai-customer-support-tools.md)
- [Customer Feedback Analysis with AI](https://cotera.co/articles/customer-feedback-analysis-ai.md)