| Add-on | AI agents - Advanced |
An AI agent can relieve your customer support team of many repetitive tasks. Nonetheless, there are times when complex, urgent, or sensitive queries need to be transferred to human agents. For those situations, you should develop an escalation strategy and design the escalation flow before launching your AI agent.
This article contains the following topics:
Deciding your escalation strategy
As part of your escalation strategy, you should establish the rules for which intents should be escalated and through which channel. Typically, there are three channels of escalation: in-chat, ticket, and email.
Since this affects agent handling time, there are a few things to take into account when deciding the escalation strategy, including:
- Size of your team
- Volume of your chats
- Agent office hours
- Complexity and urgency of the query
For example, if you are in fashion e-commerce and your team has 3-5 agents, working 9am-5pm, with a chat volume of 200-250 chats a day, escalating everything would result in long waiting times for the customers, and overload your agents.
So, in this example, it'd be good to reserve in-chat escalations for issues such as sizing and product queries or delivery failures, as these issues might affect revenue or cause a high level of frustration that might impact CSAT. Typically, offer escalation only when there's no possibility that the AI agent can guide customers to self-service.
Designing the escalation flow
After deciding which intents require in-chat escalation, think about what the AI agent can do prior to escalation, including:
- Gather necessary information that could help with handling time, such as order number, name, or email
- Add tags and fields to be added or updated specifically to your workflow
- Identify suitable agents or departments to take the chat
After those steps have been taken, you can design an escalation flow in template replies and link the dialogues that require escalation to it. If more than one escalation flow is needed, you can create more templates.
Building escalation paths
An escalation block allows you to plan how a customer will be transferred to other support channels if the AI agent can't resolve the issue.
Each escalation path contains:
- AI agent message to communicate to the customer what is happening.
- Availability block to check operating hours and team availability.
-
Escalation methods
- Send an email: Enter the email address of the agent or support department where the ticket should go.
- Forward to an agent: Select the department for the escalated chat. The default department is set in the integration parameters for the chat platform.
- Custom escalation: Add a label for analytics purposes to mark escalation traffic to another location, such as web form or hotline. This method does not escalate toward a specified endpoint.
0 comments