Here are support ticket summaries for my 20 highest-value accounts over the past 3 months. For each account, I've listed: ticket count, brief topic summary, and tone notes (if any).

[Paste your account-by-account ticket summaries]

Analyze these accounts and give me:

1. **Churn risk ranking:** Sort all 20 accounts from highest to lowest churn risk. For the top 5, explain your reasoning in one sentence each.
2. **Pattern detection:** Are there common themes across the at-risk accounts? (e.g., same feature area, same onboarding gap, same team size)
3. **Silent risk:** Which accounts have suspiciously LOW ticket volume? No tickets can mean satisfaction — or it can mean they've stopped trying and are quietly evaluating alternatives.
4. **Intervention playbook:** For each of the top 5 at-risk accounts, suggest a specific outreach action — not "schedule a call" but a targeted message addressing what the ticket data reveals.

Mark any inferences about account health that go beyond what the ticket data shows with [VERIFY] — I need to cross-check these against usage data and account manager notes.