Chatbots are the new black of IT. Every customer-centered company wants to have one. They are fancy, smart and efficient. According to Gartner, by 2020 as many as 85% of organizations will have decided to implement a chatbot. This technology is successfully used in marketing, sales and HR departments, but most implementations can be observed in customer service departments. The Internet is full of stories in which the implementation of a chatbot brought about an increase in employee productivity (by 350%!), an increase in customer satisfaction (by 22%) or a reduction in the number of requests to the hotline (by 50%). Examples can be multiplied, and the popularity of chatbots has bred a lot of internet myths about customer service automation and its effects. We have chosen three that we meet most often in our work.
#1 Chatbots only automate simple requests
Most chatbots available on the market actually only support simple commands and answer basic questions. Their high popularity is associated with low price and fast implementation, unfortunately at the expense of functionality. Process-based bots lack NLP (natural language processing) features. This means that the level at which they can understand users’ messages is very low, but sometimes it is enough for the customer to be satisfied with the service.
But this is a myth. There are many chatbot solutions that can also handle complex queries. Conversation-based bots, which use NLP technology, are able to understand what the customer has typed in and react accordingly. They deal perfectly with the context of the conversation and synonyms, and speaking with them is a lot like having a conversation with a real person.
#2 It takes a week to implement a chatbot
Platforms such as ManyChat or Chatfuel allow users to create and implement process chatbots on their own, e.g. using Facebook Messenger. No programming skills are required for this, and the process itself can be done really fast – even within a week. A chatbot built in this way will allow you to automate simple processes or answer basic, frequently asked questions.
So, is this assumption correct? The answer is: it depends on the target and type of bot. A process-based chatbot can be successfully implemented within a week and bring very good results to your customer service department. However, if your business needs a chatbot that is more advanced, understands natural language and learns from real conversations conducted by consultants, it will take more than a week to implement it.
I wrote about what kind of chatbot to choose for your Customer Service department in my previous article. Read it to find out how much time you need to implement a chatbot in your customer service – link.
#3 Chatbots will replace consultants
The aim of implementing a chatbot in a customer service department is to reduce the costs of running the department and increase the level of customer satisfaction. According to IBM, bots are able to respond to 80% of the most common requests. Automation of the most frequent ones will relieve your consultants and restructure the department in such a way that only cases requiring human intervention will be actually handled by humans. Such a strategy reduces the number of jobs and speeds up the development of employees’ competences. So is it a myth, or is it true?
It is true that bots will replace consultants in jobs that are repetitive and involve answering typical questions. However, they will not replace consultants completely – there will always be cases where human intervention is needed. There are solutions that enable developing the bot’s skills during conversations conducted by human consultants. The percentage of bot coverage is constantly increasing, which leads to lower Customer Service costs by reducing the number of employees.