Summary

Want your AI Support Agent to take smarter actions based on customer behavior? This guide explains how to set up automated rules that help the AI detect specific situations, like when a customer asks for a human, when the AI can’t answer a question, or when a chat is successfully resolved. You'll also learn how to configure the right actions, like routing chats, sending custom messages, or updating the chat status.

Instructions

The AI Support Agent comes with 3 pre-defined rule types.

Note: These rules are turned off by default, but you can edit or enable them as needed. Once active, they allow the AI to automatically respond to different chat scenarios.

Available Rules for AI Support Agent

Here are the three rule types you can configure:

1. Detects when a customer wants to talk to a human

This rule identifies when a customer expresses a desire to speak to a human agent. Phrases like:

“I want to speak to your support executive.”
“Connect me to a human agent.”

...are recognized as triggers. Once detected, the AI will take the action you’ve set up, like routing the chat to a team.

2. Detects when the AI cannot answer a query

This rule activates when the AI Support Agent doesn’t have the answer. For example, if your AI is trained only on Wati-related topics but the customer asks a question about Facebook, the AI can’t respond confidently. Instead, it follows the action you’ve configured, such as sending a fallback message or transferring the chat.

3. Detects when a chat has been resolved

In this rule, the AI uses customer responses to figure out whether their issue has been resolved. Positive or confirming messages from the customer help the AI make this decision. Once a chat is marked as resolved, the AI can take follow-up actions like updating the chat status or tagging the conversation.

Actions the AI Support Agent can take

Once a rule is triggered, the AI Support Agent can carry out different actions:

Route chat: Send the chat to a specific team or team member.
- Business plan users can use Round-Robin assignment when routing to a team.
- You can also set a custom routing message. For example: "Sure, I understand you’d like to speak with our support team. I’m now passing this chat over to our customer support executive, who will help you further."
- After routing, the AI Support Agent will no longer respond in that chat.
Start chatbot: Trigger a specific chatbot from your list.
- Choose the chatbot you want to run.
- After starting the chatbot, the AI Support Agent will no longer be active in that conversation.
Send a custom message:
- Use this when the AI can’t answer a question but should stay on the chat.
- You can define what message the AI sends to the customer in these cases.
Update chat status:
- When a chat is resolved, the AI can update the chat status to Solved or Pending.
Add tags:
- Automatically tag chats in Team Inbox.
- Ideal for categorizing chats that the AI has marked as resolved.

How to set up a rule for AI Support Agent

Follow these steps to create or edit a rule:

1. Open the rule editor:

Click the edit icon next to an existing rule
Or click Create New Rule

2. Choose a trigger:

From the dropdown, pick what should trigger the rule (e.g., “Customer asks for human”)

3. Add a filter (optional):

Click Add Filter
Select a specific AI Support Agent or leave it blank to apply the rule to all AI Support Agents.

4. Set the action:

Click Action, then choose an Action Type (e.g., Route Chat, Start Chatbot, etc.)
Fill in the action details, such as choosing a chatbot or writing a custom message

5. Save and enable the rule:

Click Save changes
Switch the rule ON to activate it

By using these rules, your AI Support Agent can handle chats more intelligently and ensure a smooth support experience for your customers.

Frequently Asked Questions (FAQs)

General Questions

1. What are AI Support Agent rules?

→ These are automated rules that help the AI Support Agent respond to specific chat scenarios, like when a customer asks for a human agent, when the AI doesn’t have an answer, or when a chat is resolved.

2. Are the rules enabled by default?

→ No, all rules are turned off by default. You need to manually edit and enable them to activate their functionality.

3. How many rule types are available?

→ There are 3 rule types:

Detect when a customer asks to speak with a human.
Detect when the AI Support Agent can’t answer a question.
Detect when a chat has been successfully resolved.

Rule Setup for AI Support Agent

4. How do I create a new rule?

→ Click on Create New Rule in the rule editor. Then, choose a trigger, apply filters (if needed), set the desired action, and click Save changes. Finally, switch the rule ON to activate it.

5. Can I apply a rule to a specific AI Support Agent only?

→ Yes. In the Add Filter step, you can select a specific AI Support Agent. If left blank, the rule applies to all AI Support Agents by default.

6. What triggers can I choose for a rule?

→ You can choose from triggers like “Customer asks for human,” “AI is unable to answer,” or “Chat is resolved.” These are selected from a dropdown during rule setup.

7. What happens when the AI routes a chat?

→ The chat is sent to the selected team or team member. If you’re on the Business plan, you can enable Round-Robin routing. Once routed, the AI Support Agent stops responding in that chat.

8. Can the AI send a custom message before routing a chat?

→ Yes, you can set a custom routing message to let the customer know they’re being transferred. This message is optional.

9. What is the “Start chatbot” action?

→ This action allows the AI Support Agent to trigger another chatbot from your list. After the new chatbot starts, the AI Support Agent stops participating in the chat.

10. Can the AI stay in the chat after it sends a custom message?

→ Yes. If you choose the Send a custom message action when the AI can’t answer a question, it will remain assigned and continue handling the chat.

11. What does “Update chat status” do?

→ It allows the AI Support Agent to change the chat status to Solved or Pending, usually after detecting that a conversation has been resolved.