ACADBOT

A chatbot to guide incoming students and connect them with current students .


Interaction Design | Conversational design

Acadbot mockup
 


Problem space

New students at universities, especially international students, face difficulty in getting information about academic opportunities, courses and part-time positions, and getting in touch with current students before coming to the campus.


Duration: September - December 2017

Work setting: Individual project

 
 

SOLUTION

A chatbot to help new students at University of Michigan School of Information access academic information and connect them with current students in specific roles. 


Goals

  • Increase accessibility of academic information for new/ incoming graduate students at University of Michigan School of Information

  • Simplify the process of finding and connecting with current students in specific roles


Constraints

  • Ensure that privacy of current students is not breached

  • Don’t overwhelm current students with questions by incoming students

  • Limit to School of Information to narrow down scope of problem statement

process

 
Process.png
 


What is the current scenario?


Comparative analysis

I analysed both direct and indirect competitors (other tools that might be used to bridge this information gap) such as LinkedIn, Facebook groups, University of Pennsylvania's incoming student portal, Canvas and Piazza (tools used at University of Michigan). I compared the features present in each platform and gauged the requirement of those features in my solution. Some of the gaps I identified included lack of availability of complete academic information and updated student information.

 
 Comparative analysis of direct and indirect competitors

Comparative analysis of direct and indirect competitors

 


What kind of academic information do students want?


User interviews

I involved target users in every step of the design process through participatory design. I conducted interviews with six participants, which consisted of 2 international students who were recently admitted to different universities in the US, 2 new students at UMSI, and 2 current students at UMSI. The aim was to understand general problems that new students face in getting access to academic information, find gaps, and understand current interaction among students.

 
 
 
 
 
AcadbotInsights.png
 
 

Personas

Based on user feedback collected from interviews and preliminary research, I created three personas which reflect the goals, needs, frustrations and motivation of the target users and stakeholders, and one anti-persona to illustrate the audience that I was not designing for.

 
 
 
 



How might we enable new students to access academic information and connect with current students?


Brainstorming and sketching

I ideated and sketched eight possible solutions to the problem I was trying to address. I wanted to be creative and get as many ideas as I could, both physical and digital, which is why feasibility took a backseat for this exercise. Out of the possible solutions, I narrowed down a few that were feasible, implementable and addressed most user needs.

 
 
 
 


Some of the possible feasible solutions were: 

 
 


What are the use cases?


Storyboarding

I used storyboarding to illustrate detailed paths the user would take to arrive at any of the solutions I had brainstormed previously. The idea of having a clear use case that starts with a problem, highlights the need for a solution, and culminates in satisfaction on the user’s end helped me better understand the user journey and account for problems they faced along the way. This exercise also helped me understand whether the motivation was strong enough for a user to give preference to a particular solution. 

 
 
 
 


Journey mapping

Through journey mapping, I identified the possible points of intervention, the gaps in the user journey and alternative routes for finding the same information. I realised that a lot of information was scattered across various sources, and there wasn't a need to repeat information that was already available elsewhere. 

 
 
 
 


Bridging the gaps


Final solution

The final solution I arrived at is Acadbot: A personal academic helper or a chatbot that assists students and guides them through the process of on-boarding. The reason I chose a chatbot is that it is a medium to make on-boarding personalised and less intimidating, as many students currently find it to be. Moreover, instead of providing a database of contact information, the chatbot will display contact information of students only when asked specific questions such as "Who was the GSI for SI 622 last year?" or "Who is the current officer of the organisation SOCHI?".

This would ensure that students are approached only when there is a real need, and are not bombarded with messages or irrelevant queries, as they might be if there is a publicly available database of contact information. For my solution, both mobile application and website have merits and demerits. However, I chose to design a website because opening external urls and reading long passages (such as syllabus, course timings etc.) would be easier on a website compared to a phone. Some of the critical features of this platform are:

  • Searching for academic information (for information available elsewhere, the chatbot would provide a short preview coupled with a link to the main source)

  • Frequently Asked Questions

  • In-platform messaging

  • Searching for a student in a specific role


QOC design synthesis and design defense

I identified one critical feature of the platform and outlined three different approaches towards this feature. Since the goal is to increase accessibility of academic information, the most critical feature is searching for relevant information/ people and getting results from the chatbot. I explored different ways of entering profile information, navigating and searching for information, displaying search results, saving questions and responses, and connecting with other users, subject to different criteria. 

This exercise helped me incorporate important elements in my final design solution such as suggestive text and categories of academic information, which ensure a guided and personalised experience. I also added in-platform messaging as it would help in connecting students without the need of a third-party social media platform. 

 

Listing Questions, options and criteria and analysing positive and negative assessment

 


Creating conversational flows

After mapping out critical features, I created conversational flows for the chatbot for a set of topics that students had most questions about, according to user research. Literature and guidelines for designing conversational UIs was extremely helpful in creating these flows. The categories for which I mapped out the flows are: Courses, Find a senior, Internships, Jobs, Student organizations, GSI positions, GSRA positions and Labs.  


Paper Prototyping

Paper prototyping was a quick way to test my designs and iterate to incorporate suggested improvements.Feedback from paper prototyping helped me account for part-time students (by adding a 'start and end year’ field instead of ‘year of study’ in the ‘Enter Information’ screen) and for people who were still unsure about which mastery course they wished to take (by adding 'undecided' as an option to ‘Choose Mastery course’ dropdown). I also added a ‘Courses taken’ and ‘GSI/GSRA/Leadership role’ text input field (optional) to the ‘Edit Profile’ section. Since courses taken by students are not public information, filling out this section would make finding peers who have taken a certain course easier.

 
Webp.net-gifmaker (1).gif
 


How do users navigate?


Wireframes and user flow

I mapped out the user flow and created wireframes before making high-fidelity mockups, wherein I incorporated the feedback from the paper prototypes. The main pages are Login, Enter information, Chatbot home page, Chat with a student, FAQs, Search, Profile and Edit Profile. 

 
 

System overview

 
 


Digital Prototype

To create the high-fidelity mockups, I studied the interfaces of existing chatbots and messaging platforms such as Google Allo, Slack, Messenger and WhatsApp. Also, since talking to the chatbot is the highlighted feature, I did not create an additional 'Categories' page. Instead, I incorporated categories within the chat to prevent the need of navigating away from the chat. 
 

Link to interactive prototype 

 
 
 
 
 
 


Reflection

I followed a structured, comprehensive design process in this project focussed on iterating in the low-fidelity stages. This was particularly helpful because the feedback led me to change my design solution from a student database to a chatbot to better address student needs. One challenge that I faced was that due to limited user research, I could not include conversational flows for all courses. Another issue was determining the scope in terms of whether the platform should include alumni, just first years or second year students as well. Currently the website is for first year students (primary audience) and second year students (secondary audience). In the future, alumni can be added on the platform to create a strong network within the University.