Managing Escalations

[Introduction]

Now let's talk about a critical part of AI powered support system. Managing escalation. We all been there. Frustrating by chatbots that loop handedly or act like a brick wall. Then all we want is to talk to a real person. You're stupid. But with smart AI system like customa escalations are roadblock. They're a feature. When aura can fully assist, it hands things over to your human team smoothly and without the drama. That's how support should be. simple, seamless, and frustration free. Let's dive into why escalation happen and how to handle them like a pro. What is an escalation?

[What Is an Escalation?]

An escalation happens when the AI decides best to involve a human agent. And here's the thing, it's not a failure. It's a way to ensure customers always get the help they need, even when the AI doesn't have an answer. Escalation are all about avoiding dead hands and delivering a great customer experience. Those are the six reasons

[The 6 Reasons Aura Escalates]

for escalation. These are the few key reasons why aura might escalate. Let's break them down. First, human requested. Sometimes customers just want to talk to a real person. Look, robot, robot, robot, robot, robot, robot. Plain and simple. If they ask to speak with someone, Aura will escalate immediately. Second, low confidence. When Aura isn't sure about an answer, it's better to escalate than risk giving an incorrect response. Third, missing information. If the knowledge base doesn't cover a topic, aura knows when to step aside and let a human take over. Fourth, user redirection. Sometimes customer change topic need conversation. If the new topic isn't covered by the current flow, aura escalated to uh the team to avoid confusion.

Five, service error. If something goes wrong like OpenAI service outage, Aura escalates so your team can step in and handle it. Six, loop detection. or is smart enough to detect else loop caused by autoresponders. For example, when it spots one, it escalated before things spiral into chaos. Those are the best

[Best Practices for Managing Escalations]

practices for managing escalation. Knowing why escalation happened is great, but managing them effectively is where the magic happens. Here's how to keep things running smoothly. First, create a dedicated chat flow for escalation. Set up a flow that routes escalation to the right person or team based on rules or internal routing. This keeps the process efficient and frustration free for the customer. You will learn how to do it in the third module. Second, use assignment rules for specific topics. For example, send building related escalation directly to the your building team. This ensure customer connect to the right expert quickly. Third, send a custom acknowledgement message. When an escalation happens, let the customer know with a friendly message.

If you do support multiple languages though, set up flows to ensure customers get these messages in their preferred language. Fourth, track escalation reasons. Use data to identify trends. If many escalation happens due to missing information, maybe it's time to update your knowledge base. Or if low

[Wrap-Up & What’s Next]

confidence keep popping up, review and refine your AI training. Wrapping up, escalation ensure customers never at that hand or escalate for six key reasons. human requested, low confidence, missing info, user redirection, service error, and loop detection. Now, by managing escalation toughfully, you will create a support system that feels effortless for your customers and deliver the human touch when they need it. What's next? In the next lesson, we'll dive into setting up tones and formality levels. A crucial step to make sure your AI sounds like it belong in your brand and resonate with your customers. See you there.