How Smart Agent Allocation in Qiscus Multichannel Chat Can Handle Your Customers’ Needs

In the case of modern customer service, each of single customer has their own query and needs. As businesses, you will need to handle all these queries based on the customer’s needs. Most of the time, the answer to queries comes from many people and various departments of the organisations. Hence, ideally, a customer’s problem should be handled by an agent who has the expertise to answer to that specific need. In the world of conventional customer service, all customer queries will come to each customer service agent randomly and will then be redirected to another agent or personnel who can answer the question. This is a cumbersome process which requires the customer to repeat the problem multiple times.

Ideally, the customer service system should understand the problem that a customer is facing and then smartly allocating the right agent to assist him or her. The agent also needs to see the chat history in order to understand the context right away without the customer having to re-explain the problem.

qiscus multichannel customer service chat

By using the Qiscus Multichannel Customer Service Chat (Multichannel Chat), you can integrate your customer service system using a chatbot that can handle and ask customers preliminary questions, and then integrate the chatbot to Rest API of Multichannel Chat that is specific to the Agent Allocation need.

Firstly, you will need to integrate Multichannel Chat with any chatbot that you prefer to use. For further instructions on how to do this, you can refer to this manual.

Subsequently, you can use the chatbot to understand what is the intent of the conversation. Once the bot understands this, you will need to program the bot to call Multichannel Chat API in order to allocate the specific agent into that conversation. You can see the list of the API needed in order to carry out agent allocation here.

The above two steps are all you need to create awesome and automatic experiences for your customer service system to execute Smart Agent Allocation.

How Smart Agent Allocation Works

The default Multichannel Chat Agent Allocation is carried out by measuring the number of activities and the load that each agent is handling. The system will automatically allocate any new customer to the agent who is the least busy. However, Smart Agent Allocation in Multichannel Chat is customisable. This means you can define your own rules and even allows you to implement AI to determine how the agent allocation works.

To activate custom Smart Agent Allocation, you will need to turn on Custom Agent Allocation:

qiscus multichannel cs chat

From the steps in the instruction above, you will need to put your own service URL to get the event whenever there is an upcoming new request. You can use this as reference in your own backend system regarding each customer activity.

The webhook for agent allocation is as shown below:

{
  "app_id": "oni-xxxxxxxxx",
  "source": "qiscus",
  "email": "[email protected]",
  "avatar_url": "https:\/\/d1edrlpyc25xu0.cloudfront.net\/kiwari-prod\/image\/upload\/75r6s_jOHa\/1507541871-avatar-mine.png",
  "extras": "{\"a\":\"s\"}",
  "room_id": "123456"
}

For further steps, you can leverage on your chatbot or any activity in your app that triggers the customer’s need to be handled by a specific agent.

For example, if the customer specifically asks the chatbot regarding payment matters, and if you want a certain agent to handle this query (especially if you want any conversation regarding payment to be handled by a human), you can make the bot to call the Assign Agent to Room. This will allow the agent you called to enter that specific room, handling the rest of the conversation regarding payment.

Another example: if you find that the bot cannot understand conversations – by the indicator that the confidence level intent is very low (less than 50% confidence level) – the bot will then call a specific agent to come into the conversation by calling Rest API (Assign Agent to Room). The agent will then be able to come into the conversation as a human to answer the challenging or complex question.

More advanced examples for smart agent allocation can be used in the market place or e-commerce. Let’s imagine how it works in an offline traditional store or mall:

  • You will come to the store that sells many variants of clothes, shoes, bag, etc.
  • You are looking at a certain product, let’s say shoes.
  • While you are looking at the shoes for a while, there will be shop assistants who would normally know a lot about shoes, coming to you and then offering you some shoes that you might be interested in.
  • This specialised shop assistant who understands a lot about shoes will convert them into real deals.

How can we do this in an online store? Using Qiscus Multichannel Chat Smart Agent Allocation, you can do the following:

  • Firstly, you will need to put an event handler in each of the product section.
  • After that you will need some listeners in the backend to that event,
  • You will then need to send automatic messages to customers offering some kind of product related to the ones they are looking at.
  • After they respond to that specific chat, we can configure the system to invite a specific agent who understands about the product to come into the conversation to attempt to convert them into real deals.

There are many more varied examples you can implement using Qiscus Multichannel Chat Smart Agent Allocation. If you are interested in implementing into your company operations or want to understand more about this, you can contact us at [email protected].

Comments

Leave a Comment

Show Buttons
Hide Buttons