Templates, Sequences & AI Agent

Everything that controls what the AI agent says, when it acts, and who gets notified.

📤 Outreach Templates

RESPOND = triggers AI reply · IGNORE = never process

Type

Description

Match Pattern (unique substring from the outbound message — no name placeholders)

Full Body (optional, for reference)

✓ Accepted Patterns → Auto-Sequence

Inbound messages matching a checked pattern trigger the sequence · the others go to the AI agent

— selected

Loading…

⏱ Sequence Steps

Two separate sequences depending on what the lead did. Any reply cancels the remaining steps and hands the conv to the AI.

Step 0 → immediate | ...

🤖 AI Agent Settings

Edit knowledge, system prompt, model and escalation recipients · changes are live, no restart needed

⚡Every 2 min

🤖 AI Agent Master Switch

Group-level AI auto-reply toggle. When DISABLED the agent stops replying entirely for this group (pipeline + booking-cron keep running). Use the per-profile pause in the left sidebar of /index.html to stop individual profiles from sending anything automated (drip, sequences, reminders) — that covers all flows at once.

🤖 AI Agent auto-reply · ACTIVE

Runtime

Model

API key loaded from .env · provider auto-detected from model name

Email recipients

Escalation (forward_to_dan)

Who gets notified when the agent forwards a conversation for human review

Daily report

Who receives the end-of-day revenue summary

Gifting export

Receives the daily xlsx of newly accepted LinkedIn leads with Apollo email. Leave empty to disable.

Gifting send hour (Europe/Rome)

Hour of day (0–23) at which the gifting xlsx is built and emailed. Default 18 (6 PM Rome).

Gifting chat message (sent on LinkedIn after the xlsx is dispatched)

When the daily gifting xlsx is built, each lead included in the file also receives this short chat message on LinkedIn from the profile that owns the conversation. Leave empty to skip the chat send. Supports placeholders: {name}, {first_name}, {email}.

AI Follow-up Settings (when lead goes silent after AI replies)

After the AI replies and the lead goes silent, the AI composes and schedules personalized follow-ups. After all follow-ups without a reply → auto close to Snov.

First reminder after (hours)

Hours of silence before the first follow-up is scheduled

Number of follow-ups

How many AI-generated follow-ups to send

Spacing between follow-ups (hours)

Hours between one follow-up and the next

📣 Campaigns (Instantly & Snov)

Choose which campaign receives this group's data. Instantly = the never-replied pipeline (Apollo → Instantly). Snov = the replied-then-silent handoff. Roles are fixed — you only pick the destination. Add a new one by pasting its dashboard URL.

Instantly campaign · pipeline (never replied)

Snov list · handoff (replied then silent)

➕ Add a new campaign (paste URL)

System prompt

The behavioral brief the agent receives on every turn. Kept outside the knowledge base so you can tune voice without touching facts.

Knowledge base

Markdown source of truth. Stored at knowledge/{group}.md. Cache invalidated on save.

⭐ AI Examples

Winning conversations the agent reads on every turn — learns tone & pacing from real wins.

💬
Lead message

→

📋 System prompt
Rules + Knowledge + ⭐ Your examples

→

🤖
LLM turn

→

✉
Reply to lead

Currently active for Bookspert 0 / 5

Find conversations to add

🔍

Booked only

0 selected

Templates, Sequences & AI Agent

📤 Outreach Templates

✓ Accepted Patterns → Auto-Sequence

⏱ Sequence Steps

🤖 AI Agent Settings

👁 How the AI sees your examples