AskIT Chatbot

The AskIT chatbot focuses on creating an interactive platform for New York University students, staff, alumni as well as other categories of people to get answer for their questions on various functional areas across various locations of NYU.


Experience & Interaction Designer, UI Designer

Project Type

Chatbot web app


Sketch, Dialogflow, AWS, Vanilla Javascript, Nodejs, Elasticsearch


Praphulla Bhawsar, Adeola Uthman, Matteo, Yeonhee Lee, Sarth Desai

Project Timeline

April 23th - August 20th 2018


New York University Information Technology Service Desk

The Challenge

Service Desk couldn't respond fast enough for user requests

The current NYU Service Desk Knowledge Base was complicated as the users had to select various fields and provide a search keyword before searching their questions. The problem was users usually didn’t know what fields to select before searching. Thus, searching for their questions could get tedious. They tended to contact Service Desk for petty queries which did not need customer service agent assistance. This increased traffic and might affect students having some real problems which needed a conversation with the customer service agent.

The Objective

How can a conversational interface individually and contextually communicate one-to-many?

The objective and goals of this project was to allow NYU IT Service to respond faster to user’s requests and concerns. It had to provide 24/7 services for immediate and personalized response by eliminating delays in user inquiries. And, to implement a knowledge centralization using real time responses to all recurring questions from users. 

A chatbot is a conversational interface that can communicate one-to-many individually and contextually. Individual and contextual means the interface is a communication tool to reach out a more specific person and know what the person is talking about. The contextual communication is what makes chatbots a promising revolutionary technology.

AskIT Chatbot is designed to provide an engaging communication to help NYU students, staff, and alumni find the solution to their issues without contacting Service Desk or looking at NYU Service Knowledge Base.

The Approach

Design process to create a chatbot

Team collaboration and design thinking approached were important to bring more value to the project. It helped us generating ideas to create engaging conversational interface which can be more playful than just a bot. User data, including data analytics, qualitative insights from interview and testing provided invaluable input that helped us shaped bot's personality to not only meet the project goal but make users delightful.

Designing a chatbot for NYU community

AskIT Chatbot is developed to provide 24/7 services for immediate and real-time responses. It helps two type of users to have a more efficient communication: a) NYU students, staff, and alumni, and b) NYU Service Desk staff to respond all user requests. In order to bring more value for both users, I and my team conducted user research to know in depth who were we designing for. We created two type of personas to help us designing the chatbot and help the developers visualize who were the target users.

A chatbot is nothing if it's not pleasurable for the users. To create a meaningful experience with the bot, users need to have a seamless conversation with the bot. Otherwise, users will find it boring and monotone. Designing a chatbot is really the most challenging project I've ever done. Since the experience is not just about making it visually engaging, but also how to humanize the conversations. I spent more time to create conversation flow within the context rather than the visual design itself.

Design Solution

Conversation design

Detailed Design

The chatbot personality

The chatbot personality is one of the most important points to be considered when creating a chatbot. Since the conversation is with an individual, the experience of one's comfort is directed by personality. Once I knew who our target users were, me and my team created the personality that fits with our users, in this case was NYU students.

Designed by Yeonhee Lee

Final Designed by me and Yeonhee Lee

Contextual flow

To make the bot more human and understand the context, I did more qualitative research by interviewing NYU IT Service Desk staff members to know how was user behavior in asking for help, what was the most contextual questions to ask, and how did they solve it at the moment. A mind mapping exercise helped to provide a conversation framework within the related context.

Using the mind map that was previously created, a high-level conversation flow was developed. The conversations had to match the bot personality and created not just linear conversation, but also provided many paths how users might responded. It had to include all the paths the user could take to reach the end goal. The more flexible the conversation was, the more engaging it would be. 

Writing the script was another challenge. Coming up with the right tone and feedback and how to measure the goal from the conversation had to be reflected on the script. It determined how was the best way to reach the bot’s goal through the flows.

Catching all possibilities

The tricky part of creating the chatbot was to design all possibilities from user inputs. The storylines were an important foundation to know all possible states. From there, I had to design the communication for all the paths, including how the bot would respond if user input unexpected messages.

Visual Design

Making the bot more playful

The way the bot talk represents the objective of the app. To resonate with the users, based on the previous persona which mainly was NYU students, the bot needed to establish the right relationship. Having a personality to the bot helped to create a productive human-bot-relationship and a good experience. The personality didn't only shown on the conversations, but also from visual aspects. Illustrations could engage user experience to support the conversation visually.

Illustrations made by me

Crafting Experience

Detail matters

The more human the conversation was, the more pleasurable the users would be. To achieve it, detail did matter. Micro interaction between the bot and the user would differentiate how the user felt, if user felt like talked to real person or just a bot without emotion.

Avatar and typing animation

Both the avatar and typing animation made the bot more human. The avatar brought more personality to the bot. Meanwhile typing animation created a thinking effect to let user thinks that (s)he was talking to a human.


Identified the user if the bot read the messages or not. If there was a connection issue causing the bot couldn't reply directly, it would tell the user that the messages were delivered but not read.

Color overlay

Since chatbot is a time-based platform, it would be necessary to make the user realized what was happening in a time frame. Color gradient would notify user the current and past messages.

The bot provided multiple results as user could choose, but to enhance user experience in making decision, the it offered the best match for user inquiries.


Creating illustrations for the bot would make the bot more playful and engaging.


Moving Forward

Creating a chatbot for the first time was really a big challenge. Moreover, designing a conversation that could be understand easily for college students had to align with their habits conversationally. Me and my team did some user testing to find out more about how the students felt about the conversations, was it sounded natural or not. How could we improve the bot the be more human to make the users more pleasurable talking to a bot.

Iterative design for better product

What I learned

Conversational interface is the next UX challenge

Since the chatbot era rises, users are still adjusting the transition from Graphical User Interface (GUIs) to Conversational User Interface (CUIs). As for designing the conversation, to be honest it was one of the most challenging things I've ever done. Conversation is a basic human communication tools. That said, it's natural for human to do it unconsciously. But not for machine. Like human, machine needs some time to learn the conversation. In order to make machine learn the conversation, the conversation designers have to abstract our thought patterns and unconscious habits into something that can be interacted in organic way.

Conversation is something that is not absolute. It is subjective based on the person who's interacting with. To design the conversation is different than design a visual graphic: it's an infinite loop. From my personal opinion, what matter the most when designing conversations is to always practice writing the flows and scripts. Doing observation about how the target users interact with each other is also an important thing to be done. Learning the users conversation behavior will also benefit us to write how the bot's personality will meet the goal.