Configuration
Configure Pounce's AI tone, response rules, escalation triggers, and daily limits to match your business.
Business Settings
In Settings, configure:
- Company name and description — Used by the AI to understand your business
- Industry — Helps the AI use the right terminology and context
- Website URL — Pounce can reference your site for accurate responses
- Response language — Default language for AI replies
AI Tone
Control how Pounce sounds:
- Professional — Formal, concise, business-appropriate
- Friendly — Warm, conversational, approachable
- Casual — Relaxed, informal, startup-friendly
- Custom — Write your own tone instructions (e.g., “Use sailing metaphors. Keep it under 3 sentences.”)
You can also add knowledge base entries — key facts the AI should reference in responses (pricing, hours, policies, etc.).
Response Rules
- Max AI messages per conversation — Default: 10. After this many exchanges, the lead is flagged for human review.
- Daily send cap — Default: 100 emails/day. Prevents runaway AI responses.
- Escalation triggers — Keywords that flag a lead as “hot” (e.g., “urgent”, “complaint”, “cancel”, “lawyer”). Triggered leads get a notification.
Inbox Settings
Configure inbound lead capture:
- Inbox provider — Gmail, Outlook, Resend (webhooks), or none (forms only)
- Resend — Enter your API key and the email address to monitor. Set up an inbound route in your Resend dashboard pointing to Pounce’s webhook URL
- Gmail / Outlook — OAuth-based connections. Click Connect and grant read permissions
- Monitoring mode — Real-time (push/webhooks) or polling (check every 5 minutes)
See Inbox Setup for connection steps.
Booking Settings
- Booking provider — Cal.com or Calendly
- Calendar link — Your scheduling page URL
- Auto-suggest booking — When to offer a meeting (after 2nd AI message, 3rd message, etc.)
- Booking CTA text — Customize the scheduling prompt
See Booking Integration for webhook setup.
Base Rules
Every AI response is governed by base rules — a set of mandatory instructions that apply regardless of your tone or form settings. These cover:
- Intent filtering — Automatically ignore spam, newsletters, and cold outreach
- Identity disclosure — AI identifies itself and the business
- Response boundaries — No fabricated info, no legal advice, no internal details
- Escalation rules — Immediate handoff when a customer asks for a human
- Compliance — Unsubscribe links, opt-out respect, follow-up limits
Base rules are loaded from src/config/base-rules.md at runtime. Self-hosted users can edit this file directly — changes take effect on restart. See AI Responses for the full details.
Questions about this topic?
We are here to help. Reach out or check our support resources.