Nothing to hand? Load the — notes with severity, impact, attempts and an ask all stated, and a clean escalation with zero gaps — the , where there isn't enough to escalate from and the honest reply says so, or the , where the customer explicitly threatens to cancel and the escalation has to surface that prominently.
Paste your notes — the prescan is free
No upload, no AI: the prescan reads your notes in your browser and reports word, sentence, paragraph and heading counts, and two families of flag — missing essentials (no severity language, no customer-impact count, no timeline of attempts, no clear ask) and churn/legal/PR risk language ("threatening to cancel", "legal action", "social media"). Each family explains why it matters. This part costs nothing and happens while you type.
The AI packages it — this is the metered part
A pass over your whole notes: the gaps worth filling before sending (each with a category and severity), the assumptions made to produce a complete package (each with why it was reasonable), a full escalation with severity, business impact, what's been tried, who needs to act, and the recommended next step as its own sections, and plain-English highlights of the notable choices made. For every prescan flag, the model addresses it or explains why it left it as a gap — and any detected risk language is surfaced prominently, never buried. Pricing is honest: a worst-case amount is reserved before the run and only what the run actually uses is charged — the meter next to the button shows both.
Export and iterate
Copy the generated escalation alone, or the full report as Markdown. Download the escalation as plain text, the full report as Markdown or JSON. Past escalations are saved to your SkillSafe account, so they are there on your laptop and your phone: restore one, tweak the notes, and re-run.
A derived work of @anthropics/customer-escalation (github.com/anthropics/knowledge-work-plugins).