The making of Mindbot: Exploring a student wellbeing AI chatbot project
University of Northampton AI and Data Science students Bhavya Sinha, Uttkarsha Kapoor and Angel Analyse have developed an AI-powered student mental health and wellbeing chatbot called Mindbot as part of their project coursework.
We asked them to share the journey behind the project, the challenges they faced and how it has shaped their career aspirations.
Q: Let’s start with a bit about each of you – what brought you to Northampton and what are you studying?
Bhavya Sinha: I’ve just completed my second year at the University of Northampton, studying Artificial Intelligence and Data Science. I came all the way from India, and at first, it was a bit overwhelming. Moving to a completely different country after spending 18 years in one place is a big change. But now I’ve reached a point where I understand how to handle things, how to adapt and how to go with the flow.
Plus meeting Uttkarsha and Angel has really helped improve my experience at university. We've been team-mates on this project but they're also two of my closest friends, and they've made a huge difference in helping me adjust and enjoy my time here.
Uttkarsha Kapoor: I’m also in my second year of the same course – AI and Data Science. I moved to the UK with my family and chose Northampton for university. I was lucky to meet Bhavya and Angel early on. We’ve studied, struggled and grown together. University life has taught me how important friendships and support systems are alongside the learning.
Angel Analyse: I'm a home student, originally from Bristol, now studying AI and Data Science at Northampton too. It’s been quite a shift but I’ve really found my place here with people like Bhavya and Uttkarsha. Being on the same course really helped us understand each other and naturally led to working together on this project.
Q: How did you first meet and decide to work together on this project?
Uttkarsha: I met Angel first during our orientation week. I remember looking around the classroom, unsure who to talk to, and then she walked in. She actually approached me first, and we hit it off quickly. I met Bhavya through a mutual classmate during one of the student union events. Our first interaction was a bit awkward – I thought she was a bit blunt, and she thought I had too much attitude! But over time, we grew closer and found common ground. Despite our differences, we’ve built a strong working dynamic.
Angel: Yeah, we definitely have our quirks, but what makes it work is that we understand and support each other. We naturally gravitated towards working together on projects like this.
Q: Can you tell us about the origins of Mindbot? Was it a class assignment?
Uttkarsha: Yes, it started as part of a group project module. Our professor asked us to create a chatbot for mental health support, particularly aimed at students. The idea was to provide quick, accessible help. Mindbot is meant to bridge a gap and offer instant, comforting interaction.
Q: What was your process for developing the chatbot? Where did you start?
Uttkarsha: First, we did a lot of research. We looked at existing mental health chatbots to understand what worked and what didn’t. One that really stood out to us was Wysa, and we used some of their design philosophy as inspiration. We didn’t just want to make something functional. It needed to feel personal and relatable.
Angel: To understand what students actually needed, we carried out both surveys and in-person interviews. We collected responses from 23 students across different faculties – business, science, undergrads, postgrads. Then we did about 10 to 12 one-to-one interviews across campus, which gave us deeper insight into how students feel about using chatbots for mental health support.
Bhavya: The interviews really helped. Surveys give you surface-level data, but when you talk to someone face-to-face, you can read their emotions and body language. That emotional layer helped us shape Mindbot’s tone – we wanted it to sound friendly, not robotic.
Q: So how did you turn all of that into an actual product? What was the tech you used?
Angel: We started by designing the frontend using Tailwind CSS and JavaScript, making sure the layout was simple and calming. After that, we worked on the backend using Flask with Python. We implemented things like chat history and a dark/light mode.
Bhavya: We used natural language processing for the chatbot responses. We also added anonymous session-based logging using JSON – we didn’t want to store any personal data, and that was really important to us from the start. A lot of students expressed concern about privacy during interviews, especially because the topic is so sensitive.
Uttkarsha: Technically, we all contributed across the stack. Some of us had stronger skills in design or frontend, others in backend logic and API integration. We used this as a learning opportunity to improve together. For example, none of us had used Flask much before, but we figured it out collaboratively.
Q: Once you had a working prototype, how did you test it?
Uttkarsha: For testing, we deliberately chose different people from the original interview group. That way, we could get fresh reactions. We were happy to see that their feedback aligned with what we’d heard earlier. People wanted short, empathetic responses, not long paragraphs. That encouraged us that we were on the right track.
Angel: Yeah, they liked how it felt like chatting with a friend rather than a machine. That was really important to us. Some people said it gave them just the kind of gentle encouragement they needed.
Q: Did you make any major changes after testing?
Bhavya: A few! We added a voice input feature and made some visual tweaks to make the site more engaging. We also discussed adding a “mood classifier” that would detect the user’s emotional state based on input and respond accordingly – for example, showing a sad or happy emoji. We tested it, and while it was promising, it didn’t quite work consistently, so we held off on launching it.
Uttkarsha: We’re still working on getting that right. We also kept the chatbot intentionally simple – no data collection, no complicated features. That way, it remains accessible and private.
Q: What’s next for Mindbot? Is it just a class project, or do you plan to take it further?
Bhavya: Initially, it was just for the assignment. But we got such great feedback from students and our professor that we started thinking maybe this can be more than a project.
Uttkarsha: Our professor was very supportive, and the idea of building something that could genuinely help people motivates us.
Q: And what about your own futures? Has this project changed how you think about your careers?
Angel: Definitely. I originally wanted to go into healthcare and had no background in computing before university. My A-levels were in Biology, Chemistry, Psychology. This project helped me realise I can actually bring AI into healthcare, which is a perfect blend of my two interests. I’d love to pursue that, maybe through a master's in AI.
Uttkarsha: I’ve always been passionate about technology and maths. But this experience at university is opening my eyes to how AI can apply to fields I hadn’t considered before – criminology, food science, mental health. I think I’d love to explore how AI can make a difference in diverse areas.
Bhavya: For me, my interests have always leaned more towards communication and media, even though I come from a science background. I never saw myself as a tech person, but this project gave me confidence. I’ve realised that AI and data science can blend well with communication skills, especially in building human-focused solutions. So whether I stick with tech or move into marketing, I know I want to keep creating things that matter to people.
Find out more
Read about the University of Northampton’s BSc Artificial Intelligence and Data Science course.