How to find the right tone of voice for your chatbot?

Published: Sep 25, 2019

A chatbot can pull you in and evoke specific emotions — here’s where the brand voice really matters. Building the right tone of voice defines the success of your brand. And it’s good to build this tone before you design the chatbot interface.

A chatbot can make you feel a wide range of emotions: you can laugh, you can feel sad, happy, confused or you may decide to make an action. Chatbots express human qualities through the medium of text — text that we associate with the brand.

Why your company needs a good tone of voice

Your chatbot’s tone can be predetermined or developed over time. Whatever the case: any misunderstandings created by the tone will create a negative image of your brand.

How your chatbot responds to your clients is going to dictate the kinds of clients you’ll be able to attract and retain. In other words — your chatbot’s tone of voice is going to shape your client base. Visual and text elements can deliver the right message to the right group of people. Everything matters: marketing pathways, text structure, accuracy, colours, images, fonts, logos, etc.

How to develop the right tone of voice for a chatbot?

Chatbot technology relies on voice and tone components of a conversation. An average chatbot has a textual interface where the conversation with customers appears. The tone of voice should be flexible and adaptable to the changes that happen during a conversation.

The foundation for successful communication is transmitting your values and mission to the customers via a conversation. You should be consistent — one specific thing should always lead to the same reaction from your side. You may use cheerful, sarcastic, neutral or any other tone that could fully reflect your values and mission.

You can use guidelines from MailChimp, Facebook, and other companies that will give you tips and samples of how to create your own chatbot and build the right tone of voice. But the core thing to consider is the emotional state of the customer. And we’ll tell you how to deal with it below.

The emotional state of the customer

Natural language understanding (NLU) lets you detect and extract emotions from the language of the customer. You can get it in the form of a positivity score after the user sends you a line of text. This tool can be custom and respond to the specific variety of emotions that your potential customers may experience.

You can use Tone Analyzer by IBM or make manual user research and analyze their experience to understand the states. You may end up with the following:

  • ? Confused – The state in which users are not sure which option to choose or they are not ready to make a final decision.
  • ? Frustrated – It’s a state when users cannot finish some tasks or something has come out as an obstacle.
  • ? Delighted – It’s a state when users have all their needs satisfied or all the tasks completed.

The state of the conversation

It’s good to build a shared and unified understanding for both a chatbot and a customer. When a chatbot receives the word “yes”, it should definitely know what the user is confirming. When there’s a so-called alliance within every dialogue, your tone of voice will sound the most natural way.

What can you do if there’s no such understanding:

  • Extract information from the user’s message and make a request to confirm that this information was interpreted correctly
  • Send a visual or textual message to show that the chatbot understands the user’s message
  • Send a request to get more information from the customer

For example, let’s have a look at different reactions to the word “home” sent by users. It can mean going to the place of living or to the main page of an application. Context matters and the chatbot should be able to understand the difference.

The state of the system

The system changes during the conversation is going on – it can adapt to the changes immediately or with some delays. The goal of the chatbot is to inform about the delays if there are any.

Here are the most common states of the system:

  • Error. If you have any failure of the system and the user is unhappy with it, the proper tone of voice will let you calm everyone down. You should inform about the problem and when you plan to resolve it.
  • Success. If there’s a completion of the task or the user has finished a series of steps to take, a chatbot can share this result with the customer.
  • Processing. If a task requires some time to be completed, the user should be informed about it. If it takes too long, a chatbot should organize regular notifications about the state of the process.

Text vs Voice

You can compare a textual chatbot like Facebook Messenger with a voice chatbot like Google Home in your language and see what you like more. You can check how you perceive information and which variant will match your brand the most.

Why you may like a text-based conversation. You can use visual elements that can add emotions to the plain text. You can keep records of all the conversations from the past. You can deliver visuals like photos, images, animations or even videos to transfer the right values and mission.

Why you may like a voice-based conversation. The focus of communication greatly depends on the customer. You should carefully plan and design the tone of voice to deliver the right message with the right tone.

How to set up the tone of voice?

You can always check and test off-the-shelf solutions that are available on the market or create a chatbot app from scratch that will fully respond to the needs of your business. The latter variant will let you deal with a fully customizable solution and create a unique and valuable tone of voice.

You may face the need to test some variants of the tone. And you may end up developing a few variants of voice tone that will respond to the needs of different customer’s groups. All depends on your business type and what kind of message you’d love to deliver to each customer.

right tone of voice for chatbot