WFH Tip 11: Reduce operator workload in half with Zobot actions

WFH Tip 11: Reduce operator workload in half with Zobot actions

Making sure the customer support team is functioning hassle-free is crucial for businesses around the world, especially in today's WFH scenario. The number of issues become more and sometimes there will be a barrage of requests. The task of solving all these issues can become rather challenging and time-consuming for the operators of your organization. So, even a little help matters. This post will talk about how Zobot actions can lessen the burden on your operators with respect to handling visitor queries.

What are Zobot actions?
The activities that are performed by the bot while responding to visitor queries are called Zobot actions. They accept only text as their value and the values can be one of the following - reply, forward, operator_busy, context, block, pending and end. 

  • reply : This action helps the bot handle the visitor on its own by providing replies to visitor queries. It helps the Zobot have conversations with the visitor just like a human operator does. While operators have to manually type out responses to every question from the visitor, the Zobot quickly and automatically responds to queries with built-in responses based on conversation history thus reducing the workload of an operator. replies accept only a collection(i.e. a list) as the value. The collection can, in turn, have one or more strings.

                                           
Syntax:
response.put("replies",{"Text1","Text2"});

Sample Code: 

Case 1: One or more strings - when there are one or more text responses 

response.put("replies",{"Hello there!","Welcome to Zoho SalesIQ."});

Case 2: Multiple Maps - when the response contains Display cards like articles, images, etc 

response.put("replies",{ 
{
      "type": "articles",
      "text": "Here are your reference articles",
      "image": "http://zylker.com/help/common.png",
      "description": "Articles about discounts in home needs",
      "articles": ["121212121","123434343","123334678"]
} });

Case 3: A combination of the first two - when the response consists of both text replies and cards

response.put("replies",{ "Hello there","How can I help you",
{
      "type": "articles",
      "text": "Here are your reference articles",
      "image": "http://zylker.com/help/common.png",
      "description": "Articles about discounts in home needs",
      "articles": ["121212121","123434343","123334678"] }
});
  • forward: Sometimes, when operators receive queries that are not related to their expertise, they forward it to their peers who are experts in the particular topic using the 'Transfer chat' option. The forward action in the Zobot works in the exact same way - it gives website visitors the option of switching from talking to a bot to a human operator while they are in the middle of a conversation at the click of an option. When the bot is not trained to handle a visitor, it can forward the chat to a human operator who is an expert.

                                   

Syntax: 
{
      "action" : "forward",
      "replies" : ["Your text here"]
}

Sample Code: 
{
      "action" : "forward",
      "replies" : ["Thanks for contacting us today, You will be connected to our operator shortly"]
}
  • operator_busy: Visitor queries can come in at any time of day. Those queries that come in after business hours or when all operators are busy can sometimes become missed chats because none of the operators will be able to answer them. This can make customers dissatisfied and leave the website. The 'operator_busy' action solves this issue by sending out a message with the information that operators are not available and that one of the operators would get back to the visitor shortly. This can keep visitors informed about the fact that operators are unavailable to pick up chats. In this case, the visitor will be prompted to leave a message and the chat will be marked as missed.

                                 

Syntax: 
{
      "action" : "operator_busy" ,
      "replies": ["Your text here ]
}

Sample Code: 
{
      "action" : "operator_busy" ,
      "replies": ["Thanks for contacting us today, please leave a message and our operator will contact you shortly. ]
}
  • block: Whenever operators come across a spammer/harasser's IP address and want to block it, they can add the IP in question inside the 'Block IPs' section. The Zobot action - 'block' does this automatically when it recognises a spamming IP thus saving your operators the time and trouble of manually blocking it. This action lets the bot find a vulnerability in the chat and block the IP of the visitor immediately.

                                

Syntax: 
{
      "action" : "block" ,
      "replies" : ["Your text here"]
}

Sample Code: 
{
      "action" : "block" ,
      "replies" : ["Sorry, you are blocked from accessing the chat window."]
}
  • context: The 'context' action in the Zobot saves operators the time and trouble of collecting information like name, email, etc. from visitors every time one visits the website. When the bot wants to collect multiple inputs to perform one action, it returns a context. A context is a definition of a web form that collects multiple inputs in a conversational manner and executes an action which is the context handler.

                              

Syntax: 
{  
      "action": "context",
      "context_id": "book_appointment",
      "questions": {},
}
Sample Code: 
{
      "action": "context",
      "context_id": "book_appointment",
      "questions": {{"name": "movie","replies": {{"text": "Ok, Lets make a booking with your sales execute !","icon": "https://zylker.com/meeting.jpg"},"Choose the department you would like to book appointment for ?"},"suggestions": {"Books","SalesIQ","CRM"}},
      {"name": "show_timing","replies": {"Thats a great movie to watch !","Choose the date for the show ?"},"input": {"type": "calendar","time": "false","tz": "true"}}}
}
  • pending: Generating a report or creating a profile for a visitor during an ongoing conversation can take a while and during this time, operators need to make sure that the visitor doesn't become impatient and leave. For this, they need to keep them hooked with a message or ask them if they want help with something else in the meantime. The 'pending' action in the Zobot does this automatically when a time-intensive action is in progress, thus keeping the visitor pinned to the conversation.

                             

Syntax: 
{    
      "action" : "pending",    
      "replies" : {"Your text here"}
}

Sample Code: 
{
      "action" : "pending",
      "replies" : ["Please wait while I generate a report for you"]
}
  • end: Whenever a visitor on the website has been idle for a while, operators can end the chat by clicking on the 'power' button inside the operator chat window. The 'end' action in the Zobot When the Zobot wants to end an ongoing conversation with a visitor, the end action can be used.
                              


Syntax: 
{  
      "action": "end",
      "replies" : ["Your text here"]
}

Sample Code: 
{
      "action": "end",
      "replies": ["Thank you for contacting us today."]
}

We have Sample Scripts that use all of the actions mentioned above. So, what are you waiting for? Start building your own Zobot right away!
For more information, check out our Help Documentation.

Regards,
Michelle