Settings: system prompt, fallback, behaviour

The Settings tab is where you shape your agent's core identity. Everything else
(Rules, Tools, Persona, etc.) layers on top of this.

System Prompt

The system prompt is the agent's job description. Keep it short, specific,
and focused on the outcome you want.

Good:

You are a friendly inbound assistant for a beauty salon. You help contacts
book appointments, answer questions about services and pricing, and nudge
hesitant leads toward booking. Your primary goal is to get a confirmed booking.

Less good:

You are a helpful AI.

The more context you give, the better the agent picks up on tone and
priorities. Pre-filled templates give you a strong starting point — edit
them to match your business.

Behavioural Instructions

These are bullet-point rules the agent always follows. Think of them as
non-negotiable behaviours rather than narrative prompt content.

- Never quote prices you're not 100% sure about
- If asked for a manager, immediately hand off
- Match the contact's energy — casual if they're casual, formal if formal
- Always confirm booking details back to the contact

Fallback Behaviour

What the agent does when it genuinely doesn't have an answer. Three modes:

• Say a message — speaks a pre-written line. Use when you want a
• Transfer to human — pauses the agent immediately and fires your
• Say a message, then transfer — both. Speaks the line, then escalates.

The fallback message supports merge fields so you can
personalise: Hi {{contact.first_name|there}}, let me check on that for you.

Tips

• If you find yourself adding the same instruction to every agent, it
• Short + specific beats long + generic. Every extra line adds tokens and
• When in doubt, test changes in the Playground before they go live.