4. AI-supportagent

Burhan Kesapli
  • Updated

General information

An AI Support Agent acts as a digital colleague that helps handle incoming cases via phone or chat.
It can answer questions, resolve simple cases, or transfer to a human agent when needed.

The AI Support Agent can be used in both call groups and web chats, and can be connected to a knowledge base to access relevant information and answers.


AI Support Agent in call groups

When the AI Support Agent is activated, it can answer incoming calls based on the engagement rules you define.
It can handle all calls, a certain percentage, or only when the call group is closed.


How to activate the AI Support Agent

1. Open Organization settings

  • Click the gear icon and select Organization settings.
  • Go to Telephony system and click Call groups.
  • Select the call group where you want to activate the AI Support Agent.
  • Scroll down until you find the AI Support Agent section.

2. Activate the AI Support Agent

  • Click the AI Support Agent menu.
  • Toggle Activate AI Support Agent on.
  • The AI Support Agent will use the pricing plan defined under
    Organization settings → AI → Licenses and usage.

3. Select an AI agent

  • Under AI Agent, choose which agent should be used in the channel.

4. Set engagement rules

In the Engagement rules section, define how often the AI Support Agent should take calls.

  • When open: choose an engagement rate between 0% (no calls) and 100% (all calls).
    Example: Each call has a 50% chance of being handled by the AI agent, meaning it’s assigned randomly.
  • When closed: toggle Engage all calls when closed if you want the AI agent to answer even when the call group is closed.

5. Add a role description

Write a clear role description explaining what the AI agent should do.
Example:

“Handle incoming calls, ask for the customer number, and try to resolve simple issues. If needed, transfer to a human agent.”

The role description helps the AI understand its purpose in the same way you’d brief a colleague.

6. Set a greeting

Specify the greeting phrase the AI Support Agent should use when answering.
Example:

“Hi! I’m an AI agent. How can I help you today?”

7. Add actions

In the Actions section, choose how the AI Support Agent should behave during calls.

For example, it can:

  • End the current call when the issue has been resolved.
  • Transfer to a human agent when needed if the call requires escalation.

Toggle Transfer to human agent when needed to enable this feature.

8. Preview the AI Support Agent

Click Preview AI Support Agent via chat to test how the agent behaves before you go live.

9. Save

When you’re done, click Save.
Your AI Support Agent is now ready to handle calls in the selected call group.


AI Support Agent in web chat

The AI Support Agent can also be used in web chats to handle customer cases, answer questions, and escalate or transfer chats when needed.
When activated, it works like a digital colleague that automatically takes new chats based on your engagement rules.


How to activate the AI Support Agent in a web chat

1. Open Organization settings

  • Click the gear icon and select Organization settings.
  • Go to Contact Center and click Channels.
  • Select the web chat where you want to activate the AI Support Agent.

2. Activate the AI Support Agent

  • Scroll down to the AI Agents section.
  • Toggle on AI Support Agent.
  • Activation costs 1 CC credit.

Once enabled, the AI agent can be assigned as an active agent for new chats.

3. Select an AI agent

Under AI Agent, choose which AI should be used for the web chat.

4. Set engagement rules

Under When open, define how often the AI Support Agent should take chats.

  • 0% means it won’t take any chats.
  • 100% means it will handle all chats.

Example: Each chat has a 50% chance of being handled by the AI agent.

Under When closed, you can enable Engage all chats when closed so the AI agent responds even outside business hours.

5. Write a role description

Give the AI Support Agent a clear role, as if you were briefing a colleague.
Example:

“Always ask for the customer number at the start of the chat, try to resolve the issue first, and escalate if necessary.”

6. Define actions

The AI Support Agent can perform various actions during chats. Choose which ones it should have access to:

  • End chat – ends chats when the customer is satisfied.
  • Escalate cases – escalates to a human colleague when needed (Escalate status).
  • Resolve cases – closes cases when resolved (Resolved status).
  • Transfer cases – moves cases to another chat, e.g. Finance team.

You can specify which chats the agent is allowed to transfer to under Allowed transfer destinations.

7. Preview the AI Support Agent

Click Preview AI Support Agent via chat to test your configuration before saving.

8. Save settings

When everything looks good, click Save.
Your AI Support Agent is now active in the selected web chat.


Tips

  • Test the AI agent in preview mode before going live – it’s the easiest way to see if the role description works as intended.
  • Start with a lower engagement rate to evaluate the agent’s performance.
  • Make sure the AI agent’s role description and allowed actions align with your internal routines.

Why the role description matters

A clear, well-thought-out role description is the key factor in whether the AI agent succeeds or not.
It’s not just about what the agent should do but also how it should behave, communicate, and prioritize in different situations.

To achieve the best results, the role description should evolve over time.
Think of it as a living document that you continuously refine based on how the agent performs in real customer interactions.

Once your AI agent is activated, take time to observe how it performs in practice and update the role description based on what works well and what could be improved.

Creating an effective AI agent isn’t a “one and done” project – it requires curiosity, consistency, and continuous fine-tuning.

But that’s exactly what makes the difference in how well the agent performs — and how great the experience is for your customers.

 

Was this article helpful?

0 out of 0 found this helpful

Comments

0 comments

Article is closed for comments.