ELIZA II: Chatbot that provides Emotional Support

Donburi·2023년 3월 7일


목록 보기

Project Description & Goal

Nature of the Project

"HRI (Human-Robot Interaction) for Wellbeing"

The project is about design thinking process focused around humans (Empathise - Interpret - Ideate - Verify). However, this time we had to build a working prototype, conduct proper speed dating to narrow down designs, and carry out usability testing.


Our Project

Our project was about building a chatbot that promotes mental wellbeing of students at HKUST by providing emotional support. We made this decision based on the Empathise, Interpret & Ideate processes.

"40% of consumers do not care whether a chatbot or a real human helps them, as long as they are getting the help they need." (HubSpot, 2017)

Project Journey

The cohort was informed about the project topic on 6th Mar. Since our group has fully experienced the design process last time, we had a clear idea of the workflow this time too.

However, there were some differences this time, so I asked Professor Ma for clarification. After the clarification, I pointed out the things we have to do differently this time, especially in light of Project 1 feedback.


8th March) This time, we had to think about how to swap our roles while having everyone's interests fulfilled.


We wanted the whole design process to be wholeheartedly rooted in user needs, just like last time. We started the Empathise process early on.



As HKUST students ourselves, we contemplated any situations we have faced where we needed help with any kind of wellbeing (physical, mental, financial, social) but couldn't get much help.


Again, this was the most prominent part of the Empathise process. I had quick chats with students at Starbucks or the Learning Commons to hear about their thoughts and stories.


8th March) After conducting a few rounds of interviews, I realised that I have to be very careful when designing interview questions. The majority of the students considered mental wellbeing the most important, so there were times where asking about past experience got too personal. I was grateful that some chose to open up, but also sorry if it hurt them in any way.


Based on the above, we carried out some of the Interpret process.



We realised wellbeing can be categorised largely into 5 categories: physical, mental, social, financial, and others (e.g. academic).


As you can see, the chatbot for mental wellbeing was in need the most. Students considered mental wellbeing the most important because they believed that mental wellbeing (or attitude, spirit) is the core that affects all other categories of wellbeing. They also mentioned how there are not many places where students can reach out for help and it is often difficult to open up due to societal views in Hong Kong.

Generate POV

We decided to work on personas & storyboards after having some initial ideas for mental wellbeing. This was mainly because we wanted to make sure we have multiple storyboards this time to explore the alternatives and conduct speed-dating to narrow down ideas & design.



This time, we did not have in-class ideation. Every member gave some interesting ideas here and there. Below is MindMap which shows some of our ideas (not just limited to but including mental wellbeing).


Eventually, we ended up with 3 main ideas for mental wellbeing - emotional support, stress management, and mental health education.

mental wellbeing

Jerry created storyboards that explore each idea under mental wellbeing. Using these storyboards, I carried out speed dating to see which ideas students like the most. It turned out that students related the most to the idea of emotional support. Hence, we have decided to build a web-based chatbot that provides emotional support for HKUST students.

13th Mar) For Project 1, we didn't have enough time to do speed dating in person so we did them online through social messengers. I found that it was much better to do it in person, and it helped a lot in narrowing down our idea.

This time, we had to swap roles. So based on members' interests, I distributed work as evenly as possible accordingly, and also constructed our presentation flow.


15th Mar) I really wanted to work on the chatbot - it was actually one of the reasons why I took this course - but had an interview in Korea and was severely packed with assignments and exams. I am very grateful that Cherry and Jason has volunteered to take up the chatbot part.

With everyone's hard work and efforts, we have put all the pieces together in harmony.


We wanted to make sure that our interpretation of user needs and scenarios are correct. So I carried out speed dating and usability testing with HKUST students by allowing them to try our prototype.


I have asked the users to try our prototype, rate 5 aspects (learnability, efficiency, memorability, reliability, satisfaction) and give suggestions for improvements. I have utilised the think aloud protocol and asked the users to talk me through their actions and decisions. This was to make sure that I do not make my own interpretation, and that users open up and give honest feedback/opinions.

30th Mar) I think performing offline verification and carrying out usability testing was really fruitful. We received a plethora of suggestions that would improve user experience, which we incorporated into our working prototype. It was great putting the concepts learnt in class to real-life use.

The modified version of our chatbot can be seen in the slides attached below. For example, we realised from usability testing that without introducing the chatbot at the start, the users could struggle to fully utilise our chatbot. We incorporated this feedback by having a pop-up screen at the start which introduces our chatbot.


Final Results


Personal Reflection

I think my second project journey was great.

First of all, we tried to improve from Project 1 feedback and past mistakes. For example, one of the feedbacks on our presentation was that we mentioned too many features instead of focusing more on the main ones. Also, we ran overtime as a result. This time, we tried to keep our presentation concise and timed it well by doing a few rounds of rehearsals.

Moreover, it was interesting to swap roles as I got to try new things. Something new I tried this time was recording and editing our demo video, and I believe it was a good experience that would help me with my video paper.

On the other hand, there were again a few points to that I could improve on. One major takeaway is that I should be very careful when designing the interview questions, especially when they could get personal.

Thank you!

Thank you so much for being on the journey with me!

Please feel free to reach out if you have any comments or suggestions :)

0개의 댓글