Once your AI Agent is trained and tested, it’s time to make it available to your customers.
This happens in Automations — the place where you decide when, where, and how your AI Agent operates across channels like WhatsApp, email, or any other channel supported in Superchat. Here, you can also control when the AI Agent should stop and, for example, hand the conversation back to your team.
How do I add my AI Agent to an automation?
You can set up your AI Agent in Automations in two ways:
Option 1: Directly in Superchat Automations
1. Click Automations in the navigation and then New Automation.
2. In the “With AI Agent” tab, select an automation template or create your own.
3. In the selected automation, add the AI Agent node.
Option 2: Through your AI Agent settings
1. Click AI Agents in the navigation and select the agent you want to use.
2. In the “Automations” tab, choose a template that defines when the AI Agent takes over the conversation.
How do I set up the AI Agent node?
Once you’ve chosen a template or created a new automation, you’ll find the AI Agent Node in your automation builder. Here, you decide how your AI Agent should behave within that automation.
1. Select the AI Agent you want to use
Pick one of your previously created AI Agents.
2. Add additional context (optional)
You can provide extra information that helps the AI Agent personalize its behavior or understand the conversation better.
This is especially useful when:
You’re pulling in data from an external system via an HTTP request or webhook trigger before the AI Agent starts.
You want to give short extra instructions — e.g., “This customer is part of our premium plan. Respond in a more personal tone.”
Adding context ensures your AI Agent has the full picture and provides precise, human-like responses.
3. Create custom exit conditions (optional)
Exit conditions define when the AI Agent should stop responding and what should happen next. This is how you stay in full control of the customer experience.
Superchat provides two exit conditions by default:
Human handover: Automatically triggers when the customer asks to speak with a human (e.g., “Can I talk to an agent?”).
Issue completion: Triggers when the AI Agent detects the customer’s issue has been resolved and the chat can be closed.
You can also define custom exit conditions to tailor this behavior further.
Each custom condition has:
A name (e.g., “Escalate to Billing”), and
A rule description — telling the AI when to trigger it (e.g., “If the customer mentions a payment issue.”).
💡 Pro Tip
Custom exit conditions act like tool calls — they let the AI Agent pause a conversation, trigger an action (like an HTTP request or assignment), and then loop back to the AI Agent once complete.
FAQ
Can I use multiple AI Agents in one automation?
Can I use multiple AI Agents in one automation?
Yes. You can use different AI Agents within a single automation, or even reuse the same agent multiple times. Many teams build a small workforce of AI Agents, each handling a specific part of the customer journey — for example, one for lead qualification, one for support, and one for follow-ups.
Where can I see how many credits are being used?
Where can I see how many credits are being used?
Go to Settings > Plan & Billing > Credits. You’ll see:
How many credits were used in a selected time period.
A split between credits used by AI Agents and Automation tasks.
For AI Agents: a detailed breakdown showing how many credits each individual agent consumed.
For Automation Tasks: a breakdown showing which automation used how many credits.
This gives you full transparency on where your usage — and your ROI — comes from.
Can two AI Agents respond to the same contact at the same time?
Can two AI Agents respond to the same contact at the same time?
No. If a contact is already in an active automation run with one AI Agent and a second automation with another AI Agent is triggered, the second AI Agent node will automatically fail. This prevents two AI Agents from replying to the same customer simultaneously, ensuring smooth and consistent conversations.
What happens if a conversation is in Trash or Spam?
What happens if a conversation is in Trash or Spam?
If a conversation is in Trash or Spam, any AI Agent node in an automation that reaches it will automatically fail. This prevents unnecessary credit consumption for irrelevant or unwanted conversations.