3 things you need to know to implement a chatbot

Are you thinking about improving your customer service experience with a chatbot or are you implementing a virtual assistant in your service desk? Here are a few things you might want to consider before launching your chatbot project.

3 things you need to know to implement a chatbot

1. Scope

What is the scope for your chatbot solution? Who will interact with the chatbot and what kind of questions will the chatbot be able to answer? Today, there is no such thing as an all-knowing chatbot. A chatbot needs a clear set of boundaries that govern its area of expertise and how it should act. For example – a travel chatbot would help answers questions about travel and helping you to schedule and book travel arrangements. It likely should not be expected to answer questions about other topics.

How many questions should a chatbot be able to handle, is another important item to consider. Typically, fewer questions will make the chatbot perform better.

2. Profile

Give your chatbot a profile or persona. This is highly important to create a great user experience. We want to talk to someone, not something. Give your chatbot a backstory, a name, age, place of birth etc. These questions will occur and you need to prepare your chatbot to answer them.

Having a profile or persona for your chatbot will also help you design the conversation flow. It will help you find the voice of you chatbot.

3. Conversation

The conversation is the chatbot. Chatbot’s and digital assistants are already among us and we are fairly accustomed to these interactions – familiar examples include Alexa and Google Assistant. These experiences set the expectations for a conversation with a chatbot and the expectations we have on intelligence intent and language understanding. When designing the conversation flow you need to think about what you want your users or customers to expect from your chatbot. This connects back to your scope and your chatbot´s profile.

Generally, a conversation flow starts with some greeting phrase and ends with a short summery.

In between the greeting and summary, you have a number of options for your flow :

  • Asking questions to get more information or to keep the conversation going
  • How to handle errors when the chatbot can’t answer a question
  • Information that gives the user what is requested
  • Excuses – no chatbot knows everything. This can also be used to steer the conversation in another direction

Don’t let your chatbot begin the conversation by presenting a lot of information. This is not a good way to design a conversation flow.

“We see a huge increase in the use of both Teams and AI from our customers. ComArounds way of connecting knowledge and an organizations collected wisdom to a bot Is an excellent way to enhance the human experience with new technologies”.

Marcus Gullberg, Director of Business Development, Microsoft Western Europe

Author of the blog article: Erik Evlinger ComAround, Product Owner and KCS Practices V6 certified.

Why you should give your chatbot real intelligence with ComAround Knowledge™

If you want to know more about how you can drive your support-cost down and give your customers support when and where they need it, please get in touch.

Find out more about our Knowledge Management software ComAround Knowledge™, as well as our training  and consulting services.

Related links:

A checklist for choosing advanced Knowledge Management software

Blog article: How to establish a collaborative culture to succeed in Knowledge Management

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *