Support escalation patterns

Support tickets are the most honest product feedback a company gets. Each pattern behind a pile of tickets is something the product, the docs, or onboarding could fix once instead of support answering it a hundred times.

Before you start

If .agents/product-ops-context.md exists (or .claude/product-ops-context.md), read it first. Use the team, owners, tools, meetings, and metric definitions it already has, and only ask for what's missing or specific to this task. If it doesn't exist, go ahead with what you're given, and suggest running product-ops-context once so the next skill doesn't have to ask.

Step 1: Prepare the data

Confirm what's in the export: date range, ticket count, fields available (category, product area, priority, account, plan, resolution, time to resolve, escalated or not). Note what's missing, since it limits what you can conclude.

Count with real arithmetic. If the data is a file, compute counts in code rather than estimating.

Step 2: Find the patterns

Group tickets by root cause, not just by the category agents chose. Read ticket text where available, because categories are often applied inconsistently.

For each pattern, classify what would fix it:

Step 3: Size and rank

For each pattern:

Rank by a mix of volume, escalation rate, time to resolve, and account stakes. Say which you weighted most.

Be careful with:

• One account opening many tickets can look like a pattern
• A spike in one week may be an outage or a release, not an ongoing issue
• Low-volume patterns from high-value accounts can matter more than high-volume ones from trials

Output

# Support patterns — [Date range] — [N] tickets

**Headline:** [The single pattern that would remove the most pain if fixed.]

## Top patterns
| Pattern | Tickets | Escalated | Accounts | Trend | Fix type | Suggested owner |

## What the tickets say
- **[Pattern]:** [one or two sentences, with a short example]

## Quick wins
- [Docs or onboarding fixes that need no engineering]

## Data limits
- [Missing fields, inconsistent categories, short date range]

Suggest owners by fix type (engineering, design, docs, onboarding), but don't promise fixes. Strip customer personal details from any examples.

Related skills

• feedback-tagging — tag ticket themes consistently
• voice-of-customer-digest — roll findings into the recurring digest
• process-documentation — document the fix so it sticks