Blog > Artificial Intelligence, Chatbots, Digital Transformation

Humanised chatbot: tips to give personality to your bot

 
Featured Image
 
  • A humanised chatbot uses artificial intelligence to optimize the communication between companies and customers, besides establishing a friendly relationship between both parties.
  • Humanisation is a key point for the success of a business strategy and involves defining a personality for the bot, looking at the tone of voice, visual representation, transparency and other features that we will address in this content.

Imagine contacting a clinic and being served by a robot that can almost instantly check available appointment times and schedule the procedure right away. Incredible, right?

This is possible with the development of a humanised and intelligent chatbot, which can solve demands from the simplest to the most complex in a practical way and without the need for human intervention.

Unlike employees, chatbots don’t need coffee or lunch breaks. Nor are they absent because of illness, holidays or even natural disasters that can put human life at risk.

IBM estimates that chatbots reduce customer relationship costs by up to 30%. Beyond financial savings, artificial intelligence can be programmed to cross-sell a company’s service portfolio or upsell the catalogue with recommendations based on a customer’s previous requests and preferences.

The strategy is on the rise and the Gartner Institute predicts that by 2022, 72% of company-customer interactions will be completely driven by automated mechanisms such as chatbots.

In 2019, 61,000 bots were developed by the 85 Brazilian companies that responded to a Mobile Time survey. Bots were responsible for the traffic of about 1 billion messages sent per month, double compared to 2017, which highlights the trend of an increasingly conversational relationship between companies and users. The data is available in the Map of the Brazilian Bot Ecosystem 2019.

What is a humanised chatbot

The chatbot is a robot that uses artificial intelligence technology and machine learning to enable an automated and efficient service, helping companies to provide 24-hour customer service.

It represents a major evolution of the relationship between people, brands and technology. A humanised chatbot uses artificial intelligence to learn the commands and processes that must be attended to.

In addition, it has a more personal approach, customised according to the brand’s identity. In this way, it is close to a service performed by a human being, but with the speed of an automated process and without compromising the productivity of employees.

Types of chatbots

There are chatbots that are based on rules and flows and others based on artificial intelligence.

The first refers to cases in which the answers are already conditioned by means of a button, as in “To consult your balance, type 1; to talk to customer service, type 2”. It would basically be the evolution of that kind of contact. A response that goes outside the established flow is unable to be answered by the bot.

In the second case, the robot is able to learn from the questions and interactions that are performed to answer them in the future, through the concept we call machine learning. According to Sinch cases, machine learning based chatbots are able to answer up to 80% of the calls by themselves.

A system that does not take advantage of all the possibilities of artificial intelligence, as in the first example we gave, results in an extremely mechanical dialogue and has a much more robotic appearance.

To have a humanised chatbot you need to develop a personality for your bot, starting with planning the following:

Tone of voice

Every brand needs to have a tone of voice, which defines the personality and the language that will be employed to talk to customers.

Each human being has a particular way of communicating, but people adapt to various situations. No one talks to another person at a gathering of friends the same way they write a scientific paper, for example.

The same goes for the chatbot: the ideal is that interactions are customised according to the brand identity and the goal of the service.

It is important to have a language similar to that used in everyday life, in a natural and less robotic way. Another strategy that brings good results is to adopt a positive and friendly language, which responds in a gentle way to consumer demands.

Speaking “the same language” as your consumer may seem easy, but it’s usually one of the main difficulties when developing a chatbot and requires continuous improvement. So analyse how your customer communicates. If your audience uses emojis, for example, your bot should also take that more informal approach.

Humanisation

When talking about chatbots, one of the first things that comes to mind are standardised interactions. Actually, the idea is to sound like a conversation between acquaintances rather than with a machine.

One of the main challenges of humanisation is to make the answers not look like the automated answers from the search engines. Therefore, avoid giving answers that are cold and very limited.

You need to make the customer understand that your company (the bot) is interested in helping them solve a certain problem. At the end of the service, asking if the problem was solved or thanking the contact are some ways to humanise the contact.

The power of the unexpected

Hearing something unexpected makes us pay more attention to what is being said. If you say “good morning,” you expect to get a simple “good morning” back, right? To receive a different response than that would be unusual.

The intention of the chatbot is that the person is encouraged to stay in the interaction, and this kind of unexpected response is a way to keep the person interested in the chat.

Siri, Apple’s virtual assistant, is a good example. She interacts to help you accomplish tasks, organise your information, and even answer questions. In addition, she is full of personality and has become known for her unexpected and humorous responses.

When talking about being sad, for example, she gives several amusing answers, such as “You can cry if you want. My aluminosilicate glass surface is tear-proof” or “I would offer a shoulder to cry on … if I had one.”

For this to happen, a mega job of human language mapping and intent configuration within the robot is required. In this way, it is able to respond according to different scenarios.

This implies that companies want to know their customers more deeply and are willing to configure the bot for all such occasions.

Visual representation

Many brands have visual representation of their virtual assistants, that is, they give a face to them in order to have a more humanised service.

This is the case of Bela, Avon’s bot created to speak with resellers and future resellers. This strategy is a great opportunity to reinforce your branding and get closer to customers.

Transparency

Just because we’re talking about a humanised chatbot doesn’t mean that the company should pretend that the service is being done by a human being, okay? In fact, being transparent and honest is essential to successful communication.

There is nothing wrong with assuming that the service is being done by a friendly and helpful robot, this should even be made clear from the start. You can start the conversation with “Hi, I’m NAME, the COMPANY’s virtual assistant. I am here to take your questions on SUBJECT. Shall we get started?”, for example.

If, halfway through, the bot can no longer respond to customer demands, the technology should be able to identify this need and notify the consumer that it is passing the contact to a specialist who will continue the service.