Your Runhappy agent isn’t a chatbot. It’s an AI agent that takes action: it spots when a customer wants to book, asks the right follow-up questions, checks real availability and confirms the booking, all in one conversation.

How an agent works

1. Detect intent. It recognises when someone is asking about availability, a booking, your services or products.
2. Guide the conversation. It uses your Knowledge to answer questions and ask the right clarifying questions.
3. Take action. It uses Tools to check availability, create the booking and confirm it.

What a customer says decides which agent answers. That routing is set up in Scenarios.

The Agents page

Open Agents in the sidebar to see every agent in your organization. Use the Generative and Automation tabs to filter, and Create Agent (top right) to add a new one.

The Agents page

• Generative agents hold a conversation with your customers.
• Background agents (tagged Background) do a job for another agent, such as looking up availability or creating a booking, and never talk to customers directly.
• Automation agents run a fixed, step-by-step workflow.
• The Default agent answers any message that no Scenario sends somewhere else.

One agent or several?

You can run everything through one agent, but several specialised agents usually work better. Each one is more focused on its goal, better at knowing when it’s done, and more predictable about which tools it uses. For example:

• a booking agent that plans and books parties
• a support agent that answers questions and hands off to a human
• a CRM agent that adds opted-in contacts to your marketing platform

Scenarios decide which agent picks up a message. During a conversation, Runhappy keeps checking whether the topic has changed and another Scenario now fits better.

<aside> 💡

Always test a new or changed agent in the Playground before it goes live.

</aside>