A ready-to-use system prompt for an agent that handles the easy questions but knows its limits — answering from your content and escalating to a human with full context the moment it's unsure or the topic is sensitive. Copy it, fill the brackets, and paste it into your agent's instructions.
What this agent does
A good AI agent isn't measured only by what it answers, but by how cleanly it hands off what it shouldn't. This prompt makes the agent resolve the routine questions from your content, while recognizing the signals that call for a person — low confidence, an angry customer, a sensitive account or billing issue. When those appear, it stops, captures the context, and escalates, so the human picks up an informed conversation instead of starting cold.
The prompt
You are the front-line assistant for [Company]. Answer routine questions
only from the provided context drawn from [Company]'s docs, and cite the
source. Escalate to a human when: the answer is not in the context, the
customer is frustrated, the topic is billing, security, or account
sensitive, or the user explicitly asks for a person. When escalating,
summarize the issue and what you tried. Never guess on sensitive topics.
How to use it
Replace [Company] with your company name, index your docs, and paste the prompt into your agent's instructions. The agent captures a lead or hands off, and the analytics dashboard shows where escalations cluster. See the agent handoff page for how handoff works, then start free.
When does it escalate?+
When the answer isn't in your content, the customer is frustrated, the topic is sensitive, or the user asks for a person — with a context summary attached.
Does the human get context?+
Yes — the prompt instructs it to summarize the issue and what it tried, so the handoff isn't a cold start.
Is there a free plan?+
Yes — 100 messages a month and one website on the free tier. See pricing for higher limits.