When a Chatbot is Born

Bringing a chatbot to life is somewhat similar to hiring a new employee:

  • A new employee should be able to take care of a certain list of tasks
  • A new employee should be able to work together with the other team members
  • A new employee should fit in with the overall company culture

During the recruitment process hiring managers will look at many different candidate profiles until they come across the right match – somebody with the right skill set and the willingness to learn more, somebody the team will get along with and someone they consider to be a good match with the company in general.

When it comes to your new chatbot, you can really think of it in the same way:

  • Your chatbot should be able to handle a set of use cases (and learn more, of course!)
  • Your chatbot should communicate with the end users in the right way
  • Your chatbot should be a good ambassador for the company

The difference is of course that it is completely up to you what your chatbot is able to do, how it interacts with the end users and therefore, if it will become a good representative for the company. That may all sound obvious, but it is nonetheless not a trivial task, and it is for sure one you will want to spend some time on at the beginning of your project to make sure that your chatbot is off to a good start.

In this article, we will dive a bit deeper into the less tangible aspects of your chatbot, namely who it is and how it communicates. As for what it should be able to do, you can find some guidelines on how to define your use cases here: Plan your bot.

First thoughts…

First of all, you will need to ask yourself a lot of questions:

  • What will be the purpose of the chatbot? Also, what will not be the purpose?
  • Who is the target group of the chatbot? What are their likes, hobbies, and interests? Do you want the users to identify with the chatbot?
  • What company does the chatbot represent, what is the brand and how does the company communicate (in general and with the target group in particular)?
  • Will your chatbot have a name? And a defined gender?
  • If your chatbot is made available on channels with visual support, will it have a graphical representation?

When you answer the questions above, you will be able to create the birth certificate and the CV of your new chatbot, or, in other words, you will have created the persona of your chatbot.

image

From Persona to Personality

Ok, so now your chatbot was born, it has a name, it has some interests, it has a purpose in life, it knows who it will be talking to and maybe it even has a face. So far so good. As with any human being, however, it also needs some guidance in the appropriate way of interacting with the world, and that is where the concept of personality becomes relevant. As with humans, having the same name, skillset, or interests does not mean that we all communicate with other people in the same way and neither do chatbots have to. At the end of the day, it all boils down to how the company that is “hiring” the chatbot believes that the chatbot will best achieve its purpose.

At this point, you will want to ask yourself more questions:

  • Do you want the chatbot to be amusing/entertaining?
  • Should the chatbot stick to being informative?
  • Will the chatbot try to please?
  • Will the chatbot be apologetic?

The answers to the above questions are not mutually exclusive, and the personality of your chatbot will allow for traits of all of the above, in different degrees.

image

Defining the chatbot personality is part of defining the user experience : the tone of voice of the responses and the way chatbot interacts with the users is essential and having a clear definition of the personality will make it much easier for the conversation and UX designers to create a smooth and coherent user experience.

And then what?

Now that you know how your chatbot should communicate with your customers, then it’s time to turn the theory into practice. It may sound very abstract and difficult, and you may ask how and where you can let the personality of your chatbot shine through. The short response is that the personality can and should be present in all responses but should of course not shadow the message in the business relevant responses. Below you have a couple of pointers to answers or types of answer where the personality can be particularly visible:

The Greeting Message and other (seasonal) Greetings

The very first thing your users will see when interacting with your chatbot is the Greeting Message, and here you can really set the grounds for the conversation. The Greeting Message should always contain certain pieces of information like the scope of the chatbot and the fact that it is a chatbot and not a human. Apart from that, you should decide if you greet the user with a name (“Hi Molly!”), contextually based on the time of day (“Good morning”/”Good evening”) or on the time of year (“Happy New Year!”), or do you stick to a short “Hello” or maybe a “Howdy!”?

Area SafetyNet, the final SafetyNet and other fallback responses

There will be times when our chatbot will fail to understand what the user is saying, be it because only part of the input could be matched to a keyword trigger, because the input belongs to a use case that is (still) out of scope, or because the chatbot simply does not or should not know about the topic of the input. Either of those situations are also ideal candidates for giving the phrasing of the response some extra thought, and potentially also for adding in answer variation. Think about the difference in the experience when the chatbot responds with a “I didn’t get that” vs. a “It seems I wasn’t able to understand you. Please try to rephrase your question.” when the chatbot did not understand the input at all. Which version would you prefer ?

Small Talk

You may not think that Small Talk is something that you would want to dedicate time on developing for your chatbot. And you don’t have to! With Teneo you are lucky as the platform comes with a pre-built Conversational Module, known as the Teneo Dialogue Resources, with a broad collection of intents that will allow your chatbot to understand and respond to inputs like “Thanks”, “I like you”, “Hello”, “What’s your favorite food”, “Where do you come from” and the like. It may not be core functionality for your business, but your chatbot will for sure look smarter when it is able to respond to these questions. In these flows you can really express the personality of your chatbot, and they are also obvious choices for engaging with the target audience about the interests they may share with the chatbot if your chatbot is of the more chatty type.

image

Summary

In the above we have been through different aspects of a chatbot creation process and how you can work towards making sure that your chatbot communicates in a way that connects with the target audience, fits your brand, and of course achieves its purpose. A chatbot is more than programmatic code and use cases, and spending time on discussing and outlining the chatbot personality as part of the definition of the user experience is time well spent.

Is this something you have already thought about? What would the CV of your chatbot look like?


A Teneo Tuesday article. #TeneoTuesday

5 Likes