top of page

<

>

ALLINONE CHATBOT

Conversational UI Case-study:

Making A Chatbot

With the rapid evolution of Artificial Intelligence, chatbots have become ubiquitous with customer experience in recent years. The global chatbot market is anticipated to reach $1.23B by 2025 and approximately 95% of customer interaction will be supported by Artificial Intelligence. From a business perspective, chatbots increase sales and customer interaction efficiency while decreasing service costs. A majority of consumers today interact with a chatbot at least once a month and many of them believe chatbots will soon become the primary source of communication for customer service.

 

The inclusion of AI and conversational UI into chatbots can improve accuracy and relevance, and eliminate congestion. With Natural Language Processing and supervised learning, chatbots are able to communicate like humans by providing contextually relevant responses.

​

Making A Successful Chatbot

 

Define Goals:

 

  • Make a chatbot version of the Allinone app which is designed to help users utilize communication and planning tools more efficiently.

  • Facilitate personal assistance by initiating supportive and informative conversations with users.

 

Create a chatbot persona:

 

It is important to understand target users prior to creating a persona. In this case, Millennials would be the primary users as they are the most likely to utilize personal assistance tools such as calendar, to-do list, email, translator, and messenger.

Introduction:

 

Introduce the name and give users an idea of what the bot is capable of. Creating human-like dialogue helps users become more engaged and comfortable. Even when users are aware that they are conversing with a bot, they tend to be more honest and relaxed when they are participating in a friendly conversation.

1_edited.png

Happy Path:

2.png
3.png

Side path/refocus/limit user freedom

4.png
5.png

Insults:

​

Receiving insulting messages are unavoidable. Instead of generating an irrelevant response such as “I don’t know,” giving a human-like answer helps users refocus and avoid frustration. 

7.png

Escalation protocol:

6.png

Ways to improve existing chatbots:

 

  • Ask one question at a time. It is important for the bot to comprehend users’ demands/commands accurately and encouraging users to answer multiple answers can create confusion.

  • Avoid asking open-ended questions. Granting users too much freedom/power can misdirect and neglect the purpose of the bot which can ultimately bring user disappointment. 

  • Constantly redirect users by giving straight-forward suggestions and asking task-oriented questions.

  • Adding emojis in dialogues can create a sense of friendliness and ease to better mimic human behavior. 

  • Respond quickly with a brief pause to imitate real chats between humans. 

  • Create a customer journey map with proven data and actual users’ navigating behavior.

  • Utilize contextual analytics by users’ navigating/usage history, ownership, etc. 

  • Incorporate machine learning.

bottom of page