Writing Chatbot Conversations: What You Need to Know— Part II

By Kristi Colleran - March 07, 2017


This is Part II in a series exploring the need to know aspects of designing and writing conversations for chatbots which happens to be a new and developing area of expertise! While the focus of Part I was on understanding your audience, Part II is all about defining who your chatbot is.

If you missed Part I you can find it here.

Bringing Your Chatbot to Life

While the bot and messaging platforms have made it relatively easy to stand up a chatbot, we are quickly realizing that making a good one isn’t trivial. A critical differentiator can be in how you craft who your chatbot is. Here are four key considerations you will need to make along with the implications they will have on your chatbot’s conversation.

#1 — Identity

Clearly defining your chatbot’s purpose is a critical first step in creating its identity. Will your chatbot be topical or task-driven?

Topical bots are typically for entertainment. They will need the traits of a great conversationalist: upbeat, witty, complimentary, highly engaging, good listener.

Task-driven bots are focused on helping users get a job done: buy a pair of sneakers, get a question answered about their account, or find the nearest store location and it’s hours. They will need to be more direct and efficient. But keeping things succinct and to the point shouldn’t mean your bot is perfunctory and stoic. These bots will need the character and personality of a successful customer service associate: helpful, friendly, knowledgeable, trustworthy, positive, goal-oriented.

If your chatbot is for a brand ensure that it appropriately represents the brand’s tone and values. After all, the chatbot is an extension of the brand, and the brand needs to be recognizable in how the bot reflects it.

What shall I call you? Believe it or not this is a very common question that users ask chatbots. I would recommend that you name your chatbot and make sure that it knows it’s name when asked to avoid an IDK (the chatbot’s response for “I Don’t Know”). If you prefer not to name your bot, ensure it can respond intelligently to users that ask what’s your name or some version of this intent, again to avoid an IDK. And if naming, select one that’s consistent with the identity and gender (more on this in #3) of your bot.

#2 — Voice

Once you have defined your chatbot’s character and personality, select a few traits that are the essence of it’s persona and use these to craft and hone its voice.

An excellent reference framework is The Four Dimensions of Tone of Voice developed by the Nielsen Norman Group. While the framework was developed based on website content, I think there is applicability to tuning the voice of your chatbot.

The four dimensions of tone of voice are:

  • Humor: funny vs. serious
  • Formality: formal vs. casual
  • Respectfulness: respectful vs. irreverent
  • Enthusiasm: enthusiastic vs. matter-of-fact

Each dimension is represented on a 3-point scale, with a neutral midpoint.

The Four Dimensions of Tone of Voice Scale — Nielsen Norman Group

Of interest for chatbots, in the research conducted by Nielsen Norman, was the measurable effects tone of voice had on users; specifically on users’ impressions of an organization’s friendliness, trustworthiness and desirability. The tones that performed best were: casual, conversational (serious — casual vs. serious — matter-of-fact) and enthusiastic.

In my experience to-date, I have found that users of chatbots typically have a sense of novelty and few to no set expectations about the experience because they are so new. By adding a bit of whimsy to your user experience, such as unexpected humorous responses or emoji’s, your bot can delight and provide a memorable experience. The degree to which you incorporate these will be dependent upon the brand (if applicable) and your chatbot’s purpose.

#3 — Gender

It’s a boy! It’s a girl! Or not? You choose! There are a lot of varying opinions about defining gender for bots. To-date the majority of the most familiar virtual assistants from Apple’s Siri to Amazon’s Alexa to Microsoft’s Cortana, have default female personas. For many this is just perpetuating traditional stereotypes. To counter this, some companies are allowing users to select. X.ai gives users the ability to choose their meeting scheduling assistant: Amy or Andrew. Other companies are avoiding the issue altogether by building gender neutral bots.

The good news for those of us designing and writing for text vs. voice bots is we can more easily create gender neutral bots since there is no audible voice to give away gender. This allows your users to decide if they are chatting with a her, him or it!

Of course, going back to your chatbot’s purpose, character, personality and your brand, there may be very good reasons for why you would assign gender. Just think it through carefully so your choice will not be bad for business.

#4 — Human-Like or Not

While chatbots can mimic human conversation and interact with humans, they are not human. A chatbot pretending to be human can create confusion for users which quickly leads to distrust; not a desirable trait in your chatbot. The vast majority of people say they want to be told whether it is a human or a bot they are chatting with. They don’t want to be tricked. Help your bot be trustworthy by being clear to it’s users that are interacting with a bot.

Another common human expectation is that computers are accurate, consistent and fast. Rather than trying to make task-driven bots seem human-like, focus on creating a bot that meets these expectations by solving users’ problems and addressing their needs more quickly and simply than existing solutions.

After all, a “human” experience is defined by how the user feels, not how life-like the chatbot is. Your challenge is to get the balance right and leave the user feeling as though they have had a human experience while successfully meeting their needs!

Stay tuned for Part III — Helping Your Chatbot be Successful. I’ll be covering topics that will have a direct impact on your chatbot’s initial and on-going success including: User On-Boarding, AI Tools, Bot-Human Hand-offs and On-going Monitoring for Improvement.



Case Study

Drive Event

Deploying Sciensio's Eventbots solution to improve and enhance a major industry trade show.

Get Case Study

We promise that we won't SPAM you.