Email Escalation Drafter

WHAT IT DOES

Email Escalation Drafter reads your email or support ticket content—subject, sender, message body, and issue description—to identify unresolved customer issues that require escalation, then generates a professional escalation reply draft. Choose your urgency level: Standard (24-48 hour resolution), High (12-24 hour resolution), or Emergency (4-6 hour resolution). Toggle estimated resolution time to set clear expectations. Toggle escalation contact to provide a direct point of contact. The AI generates a complete, context-aware draft that acknowledges the customer's frustration, explains the escalation path, commits to timely resolution, and sets clear expectations for next steps—saving you time and helping you manage customer expectations during escalations.

WHERE IT RUNS

Any webpage with email threads, support tickets, or customer correspondence requiring escalation—Gmail, Outlook, Yahoo Mail, Zendesk, Intercom, Freshdesk, Help Scout, and more. The script scans the page for conversation context: subject lines, sender information, message bodies, and issue-related content. No site-specific configuration needed; the detection algorithm adapts to the page structure automatically.

HOW TO USE

Install the script and navigate to any email or support ticket requiring escalation. Click "Generate Escalation" in the panel—the tool extracts the conversation context and sends it to the AI. Choose your urgency level (Standard, High, or Emergency). Toggle resolution time and escalation contact on or off. The draft appears in the panel. Review it, edit if needed, then click "Insert" to place it directly into your reply field, or "Copy" to paste elsewhere. Settings persist per site and globally—the tool remembers your urgency, resolution time, and contact preferences. Keyboard shortcuts: Ctrl+Shift+E to generate, Ctrl+Shift+R to regenerate, Ctrl+Shift+I to insert.

TECHNICAL MECHANISM

The script extracts email content using a cascading DOM selector strategy: it targets subject lines (h1[data-thread-perm-id], .hP, .subject), sender information (.gD, .go, [email]), and message bodies (.ii.gt, .a3s, .msg-body, [role="article"], .ticket-body, .support-message). Collected text is concatenated with labels into a structured input. The AI prompt includes the full thread content, urgency level (standard/high/emergency), resolution time toggle, and escalation contact toggle. Urgency mapping guides timeline: standard = 24-48 hours, high = 12-24 hours, emergency = 4-6 hours. The model generates a draft that acknowledges frustration, explains the escalation path, commits to timely resolution, and sets clear expectations. The response is displayed in a textarea with Insert and Copy buttons. Reply insertion uses native value setting with both "input" and "change" events for React/Vue compatibility. The script prefers WebMCP (navigator.modelContext) when available, falling back to DOM scraping when not.

THE PANEL

The tool runs inside BGPanel, a lightweight UI system built into the BotGentz extension. Drag the panel by its header to reposition it anywhere on screen—it snaps to edges and remembers its position per site. Click the header to collapse the panel into a small bubble; click again to expand. Resize the panel by dragging its bottom-right corner. Press Escape at any time to close the panel. The panel stays out of your way but is always accessible via the extension's toolbar.

PLEASE NOTE

Requires the free BotGentz browser extension and an active monthly subscription—the AI backend uses credits for each generation. One generation per click; regenerate counts as a new generation. The script reads only what is visible on the page—it cannot access iframes or shadow DOM. AI-generated escalation replies are suggestions—always review and edit before sending. No email content is stored or shared beyond the API request; all processing is ephemeral. The tool's urgency level guides the AI's timeline but actual response times should be confirmed with your internal escalation procedures. Sites change their markup over time; if detection fails, check for updates or use the tool on a page with clearly visible support content.