PhoenixRise

Building academic resilience through practice and AI feedback

Background

PhoenixRise is an app that helps university students build resilience and overcome academic burnout.

The idea grew from my own experience with academic burnout and seeing friends struggle with it, too. It inspired me to explore how we could better support students through that challenge.

I developed PhoenixRise independently, from research through design and testing prototypes with students to building the app in React Native with AI feedback.

Jump to the prototype

University students studying together, with one student looking tired from academic stress.

Opportunity

How could we help students overcome academic burnout?

Through research, I saw an opportunity to help students handle academic challenges. That way, they would feel less overwhelmed because they know how to tackle them.

A student thinking through an academic challenge at her desk.

Solution

Help students build resilience through reflection and practice.

Research on reducing burnout led me toward cognitive-behavioral techniques ^[1]. These strategies help people practice better ways to overcome difficult situations.

I turned this research into 3 features. Students reflect on positive moments, practice different scenarios, then apply what they learned.

PhoenixRise Daily Gratitudes screen encouraging students to write down positive moments.

Build positive mindset

Reflect on what they’re grateful for to focus on the good in their lives.

Why gratitude

PhoenixRise challenge screen showing response choices and supportive feedback.

Practice better responses

Choose the most helpful response and learn why it works.

Why practice

PhoenixRise AI feedback screen helping a student practice a response to an academic challenge.

Learn from AI feedback

Write your own response and get personalized feedback.

Why AI feedback

Defining the Problem

I wanted to understand how other students experienced academic burnout.

I explored discussions on Reddit to learn how it affected their studies.

A student describes struggling to restart thesis work after months of burnout.

A graduate describes continuing to feel burned out after college.

A student describes anxiety, oversleeping, procrastination, and declining grades.

Competitor Analysis

Feature comparison of Headspace, Calm, Insight Timer, Waking Up, and Smiling Mind across stress management, theory, motivation, personalization, and academic strategies.

Exploring the Solutions

I wanted to understand how to help students manage academic burnout.

I explored academic articles on reducing academic burnout and making practice more engaging.

Research finding that gamified mobile apps can support well-being and mental health and may improve motivation and retention.

Research caution that gamified interventions may encourage excessive gaming and gaming disorder.

Research describing cognitive techniques that change thought patterns and support more adaptive behaviors.

Design Goals

Based on the research, I set two goals to help students handle academic challenges.

Learn better ways to respond

Research showed how coping strategies could help reduce academic burnout ^[2]. I aimed to provide resources to help students handle academic challenges.

To achieve this, I explored articles and videos so students could choose how to learn these strategies.

Two learning wireframes: a resource list with learning filters and a video topics page.

Apply what they learn

Research also showed that learning from resources alone might not be enough ^[3]. I aimed to help students practice strategies for academic challenges.

I explored multiple-choice and open-ended questions to give students different ways to practice.

Two practice wireframes: one with multiple-choice responses and one with an open-ended answer field.

User Feedback

Students tested my lo-fi wireframes, and I discovered two ways to improve Practice.

Making academic challenges easier to approach

Students found coursework difficult during burnout and worried Practice would feel too hard. This showed that they needed something easier to start with.

Responding to different approaches

I gave students the same academic challenge and asked how they would handle it. Their different approaches showed me that feedback needed to be tailored to each answer.

Write your gratitude

User Flow

Design Rationale

Introducing burnout through storytelling

Adding a story helps increase engagement in an educational game ^[5]. This inspired me to introduce burnout through a story before explaining the activities.

The story is about a phoenix that rises to help students overcome burnout. It is introduced as a companion that guides students through the activities.

Two story screens introduce a phoenix rising from fire and meeting students.

Making gratitude easy to start

Students are more likely to experience negative emotions during burnout. So, I added gratitude writing as a way to boost their positive emotions ^[4].

I added three short prompts for students to share what they’re grateful for. If they’re unsure what gratitude means, they can open the tooltip for an explanation.

Daily Gratitudes screens show three gratitude prompts and an explanatory tooltip.

Preparing students to write their responses

Writing answers with feedback helps people remember better than choosing an answer ^[6]. I wanted students to recognize effective strategies and practice forming their own responses.

I added multiple-choice questions as a warm-up. Students then write their answers in open-ended questions to get personalized feedback.

Practice screens show a multiple-choice response with feedback and an open-ended answer with personalized feedback.

Gamification

Interactive Challenges

Completing scenarios successfully could help students build confidence ^[3]. I wanted students to start with manageable challenges before tackling harder ones.

I also designed different academic scenarios so students can practice various strategies.

Two interactive challenge scenarios featuring Viktor and Annie describing academic setbacks.

Progress & Achievement

Seeing progress can help people reach their goals ^[7]. So I wanted students to see their results after each practice session.

I paired these results with an encouraging message to recognize their effort. This aimed to motivate them to keep practicing, even after a lower score.

A progress summary celebrates completed challenges and shows duration and accuracy.

AI Development

I built AI feedback that responds to how each student answers each academic challenge.

I connected PhoenixRise to OpenAI’s GPT-3.5 Turbo through a Node.js/Express server I built. The app sends the answer and scenario to my server. The server then requests feedback from OpenAI, keeping the API key out of the app. I aim to provide supportive feedback that explains why a response might or might not help.

Matching feedback to the challenge

Without knowing the academic challenge, the AI could give misleading feedback. Therefore, I sent each student’s answer with its scenario to help the AI give more relevant feedback.

Server code sends the student's answer and the academic scenario together for relevant AI feedback.

Setting the AI’s tone

I gave the AI examples of supportive feedback to guide its tone. I instructed the AI to recognize effort and suggest a next step using simple language.

Prompt instructions set a supportive tone, avoid harsh feedback, and use simple language.

Showing feedback based on the answer

I instructed the AI to provide feedback and a hidden score for each answer. If they score 85 or higher, it explains why the answer works. Otherwise, it explains the problem and suggests a better strategy.

Code checks the AI feedback score to decide whether to praise the response or suggest improvement.

Usability Testing

Participants: 5 students (3 initial, 2 follow-up)
Method: In-person moderated
Duration: 30 minutes each
Tasks: Onboarding, Gratitude, and Practice

I tested whether first-time users could understand the app without my guidance.

I asked participants to explain the app’s purpose, write their gratitudes, and complete academic challenges. I recruited 3 SFU students through social media and in a campus study space.

I ran each session in person on my phone without explaining the app. This helped me see where students got stuck on their own. After each task, I asked follow-up questions and took notes to understand their action. I then compared their answers to find common problems and decide what to improve.

Making burnout relatable

Some students thought the phoenix was burnout. This distracted them from understanding burnout and how it affects their studies.

I replaced the fantasy story with a student’s experience before and during burnout. This gave students a clearer example of burnout and context for how this app could help.

Before and after onboarding screens: a phoenix story is replaced with a student sharing his burnout experience.

Explaining the purpose

Some students were new to gratitude and took a long time to decide what to write. Learning what gratitude is and why it matters on the same screen felt overwhelming.

I split the explanation into two pop-ups explaining what gratitude is and why it matters. This helped students understand gratitude faster, so they could start writing easily.

Before and after gratitude screens separate the definition of gratitude from the explanation of its benefits.

Making feedback more supportive

Students were confused when their answers were marked incorrect. They explained why it felt helpful to them, even if they weren’t the best option.

I revised the feedback to acknowledge their reasoning and suggest a better approach. This helps students understand why another approach could work better.

Before and after challenge screens show feedback revised to acknowledge the student's reasoning and suggest a better response.

Making feedback clearer

The AI’s headline made students think ineffective answers were correct. They didn’t realize the problem was explained in the smaller text.

I revised the headline to acknowledge their reasoning and explain the problem. This helps students see what to improve and use the hint to revise their answer.

Before and after open-ended response screens clarify when an answer needs improvement and provide a supportive hint.

Page 1 of 5. Overview The goal of the usability testing is to evaluate how well users did when doing three tasks in my academic burnout app, which includes understanding the story, writing the gratitudes, and answering questions based on the resilience scenario Some key findings: - All 3 participants find it difficult to enter their email and password because the input screen blocks it - All 3 participants found the phoenix story to be confusing and disconnected from academic burnout - All 3 participants struggled in coming up with the gratitude because it wasn’t well explained and felt broad. - 1 participant was able to answer all the quiz sessions correctly, but the rest made a couple of incorrect answers, indicating a valid reason why they think their answer feels the best response. - 2 participants dislike when they discover that the response feels like an AI talking to them rather than a human. - 1 participant finds that most of the UI buttons and font size appear small, indicating that it may be challenging for users with motor disability to interact with it and users with visual disability to read the content. - All 3 participants wrote a long prompt in the journal session. Methodology Testing Process - Format: In-person moderated usability testing - Duration: 30 minutes per session - Participants: 3 (2 are from Interactive Arts and Technology and 1 is from Computing Science - Date: 7 November 2025 - 9 November 2025 Tasks - Story (5 min): Sign up and explore the fantasy story - Gratitude (8 min): Write their gratitudes - Resilience (12 min): Complete two types of scenario-based exercises After the user has done each task, follow-up questions will be asked to learn more about their experiences based on how they performed the task. The tasks are performed in person so that it is easier for the user to access my app by using my phone. In addition, it is much easier for me to observe their nonverbal cues when interacting with my apps and also ask follow-up questions and note their answers. Key Finding from Story Question 1: What part of the story engaged you or didn't engage you? Summary: The phoenix story I envisioned does not match what the participants think. They

Page 2 of 5. expected that to make it more related to burnout experiences, based on my research rather than a fantasy story. Question 2: What, if anything, was confusing or unexpected about how the story was presented? Summary: Participants dislike the fantasy and preferred a more realistic, research-backed explanation of burnout and how this app is going to counter that. Overall Recommendation Use storytelling to educate users on what academic burnout is and find a better and explicit way on how Phoenix can help the user overcome it. Since research mentions that cognitive behavioral therapy and rational emotive behavioral therapy can help students to overcome burnout, I could change students to therapists and restructure the question as if the therapists were the ones asking them those questions during the REBT or CBT, and the user has to respond as if they were in those therapies. Key Finding from Gratitude Question 3: How did the prompts connect with what you wrote for gratitude? Did it feel connected, confusing, or not useful? Summary: Participants have a brief understanding of gratitude, but because the concept itself has lacked explanation, they have various ways of expressing gratitude Question 4: Describe any challenges you experienced when thinking of gratitude or using this screen. Summary: Without a clear explanation of gratitude, participants take a bit longer to come up with gratitudes. Question 5: Besides the gratitude content itself, what was confusing about this screen? (For example: navigation, instructions, layout, etc.) Summary: UI was generally acceptable as it wasn’t challenging to list their gratitude. The big problem is just a lack of context about gratitude and how it is connected with building resiliency toward burnout. Key Findings from the Resilience Scenario Question 6: Were there any words or phrases in the feedback that were confusing, and why? Summary: All words and phrases are straightforward to the user. Although 1 participant was confused about the “Quiz” terminology, as it gives a feeling that the user has to study before doing the questions. Question 7: Explain your thoughts on the reasoning behind the feedback. Do you agree with what it said? If not, why? Summary: All participants appreciate feedback logic, but wanted to acknowledge other approaches as they have valid reasons why they work well for them, rather than having a

Page 3 of 5. single “correct” answer. This is because the right/wrong framing may discourage users who have developed personal coping mechanisms. Question 8: Walk me through the process of completing these scenarios. What issues did you experience with how you were guided through the steps, and why were those problematic? Summary: Overall, all participants didn’t find it difficult to complete all scenarios. Only 1 believes that visual hierarchy can be improved so that people will be able to clearly focus on the scenario first, then see all the options. One mention to ensure that the stat has a meaningful purpose, as currently, he doesn’t understand the purpose of displaying time. Question 9: Walk me through the process of completing these scenarios. What issues did you experience with how you were guided through the steps, and why were those problematic? Summary: The AI feedback tone needed significant improvement. For instance, one mentions that “Oops, not quite” sounds like an error, while the other mentions that knowing that the prompt is created by an AI makes user have less emotional connection with the answer and believe that the result may be less accurate, causing less learning from their answer. Question 9: Walk me through the process of completing these scenarios. What issues did you experience with how you were guided through the steps, and why were those problematic? Summary: The AI feedback tone needed significant improvement. For instance, one mentions that “Oops, not quite” sounds like an error, while the other mentions that knowing that the prompt is created by an AI makes user have less emotional connection with the answer and believe that the result may be less accurate, causing less learning from their answer. Other Key Findings Question 10: Is there anything else you'd like to share about your experience, such as anything you would like me to improve on for the app? Summary: There are accessibility issues where the fonts and touch targets are small, cramped spacing, and poor hierarchy. One participant has to put more cognitive effort, such as scrolling to the journal and understanding why the quiz comes first before the journal. In addition, the phoenix backstory needs to be more evidence-based from my academic burnout research (Example: What is academic burnout and proven methods to overcome academic burnout). Because REBT and CBT are proven methods to overcome academic burnout, we could shift the scenario from a student in a university to a counsellor in a mental health center. If the phoenix is the mascot, it should appear a bit more in the narrative and UI. I should provide better information on what gratitude is and how it benefits the user, in addition to connecting gratitudes to the quiz and journal to help enhance user experience and therapeutic impact. Lastly, the current email input/keyboard behavior needs fixing because it doesn’t allow user to see their password and email that they wrote Positive feedback about the app

Page 4 of 5. All participants find it easy to navigate the app, and they believe the UI looks calming, making it easier to reduce academic burnout. The wording and phrases are easy to understand, giving them less cognitive effort to understand what it means, and spending more time thinking of the answer. Despite getting an incorrect answer, the feedback looks reasonable, and they understand why their answer wasn’t the best. All participants wrote long, meaningful prompts, elaborating their strategies to overcome that academic challenge. Since they really like tackling those academic challenges, they would like to see more questions to further improve their resiliency. Overall recommendation for the sign-up I will ensure to move the email and password text field when the users click on its input so that they can easily see what they type. Overall design recommendation for the story Use storytelling to educate users on what academic burnout is and find a better and explicit way on how Phoenix can help the user overcome it. Since research mentions that cognitive behavioral therapy can help students to overcome burnout, I would make the mascot explicitly as their support companion and change ‘students’ to ‘therapists’ and restructure the question as if the therapists were the ones asking them those questions, and the user has to respond as if they were in cognitive behavioral therapy. Possible flow Screen 1 (Meet Alicia): Alicia is a student who loves learning, but suddenly feels overwhelmed because of too many assignments Screen 2 (About academic burnout): Describe what academic burnout is and the symptoms Screen 3 (Calm the user and tell strategy): Describe a Ckey statistic of burnout and evidence-based research to overcome it Screen 4 (Meet Auri): Auri is a blue phoenix that will help the user overcome academic burnout using evidence-based practice Screen 5 (Recovery Journey): Describing our strategy by navigating them through the path “Gratitude → Resilience Skills → Reflection” to help users feel Overall recommendation for the gratitude I will provide clearer and better information on gratitude and how it helps to overcome burnout. I consider creating a loading screen that provides statistics to let the user have some awareness of academic burnout. Possible flow (Click on the information icon) Card 1: Describe what gratitude is. Card 2: Describe how gratitude helps burnout Card 3: Give an example of gratitude

Page 5 of 5. Card 4 (Optional): Give key things people need to know to make it easier for them to write their gratitudes Overall recommendation for the homepage Currently, the design of the gratitude looks like an interaction element, which I do not want. Therefore, I will change its appearance to make it look static. I will also change the hero image of the quiz and journal section to make it look more consistent with other illustrations I generated. In addition, I will also change the name “Quiz” into “Module” because I agree with a participant who mentioned that students are more likely to think that they need some preparation first before doing the session. I am also thinking of including subtitles to show my reasons why it is more beneficial to do the quiz first, then the journal. Overall recommendation for the quiz Rather than saying that their answer is incorrect because every participant did provide a valid reason why they think it is correct, it would be better for me to mention that their answer is good but less effective. Also, I need to ensure that all the feedback responses are leaning towards the correct answer, as some of the incorrect feedback responses mainly explain why it is incorrect, making it harder for the user to understand what the correct response should be. If I could find a more meaningful stat, I could replace it with the timer, as one participant did mention that they do not understand why they need to know how long they answer all the quiz questions. And ensure to improve hierarchy so that users will find it easier to read the scenario first, then the options. Overall recommendation for the journal I will make AI to provide a more human tone, particularly as a health counsellor. In addition, I will also ensure to change how AI makes the headline, as “That’s okay” could look like a positive message, and “Oops, not quite” could look like an error message. A better headline could be more tailored to the response, such as “It’s okay to feel overwhelmed, you are not alone”. If I have more time, I could enhance personalization where I could ask the user to write their academic challenge, and then the AI will provide the answer based on the user’s prompt. Solving students’ current challenges helps boost resiliency. I will also need to improve how AI answers users’ prompts so that it feels more tailored to their prompts. I will probably make the text area smaller to tell the users not to input too much, especially if they are in a burnout state, as I do not want them to put a lot of their cognitive effort into answering the question. Overall recommendation for the overall layout I will increase the font size and the text of the button to enhance accessibility. Make the buttons' labels more direct, but still motivational (For instance, users hesitate to click ‘Back Home Stronger’ and ‘Explore mindset exercises’ as they do not know where it will lead them) Minor changes - Fix the spelling of Phoenix as it is currently spelled incorrectly for the app name. - Change the formula, as it provides the wrong output for the quiz session - Change ‘quality’ to ‘accuracy’ as it sounds confusing

Did the changes make the app clearer?

After redesigning the app, I tested it with 2 new students without my guidance. Both related to the student’s story and found Auri’s feedback clear. However, one found it too long to read across several questions. Therefore, I plan to show the key points first, with an option to read more.

The other student felt that courses they didn’t enjoy could also cause burnout. This led me to explore a feature that connects their studies with their interests and goals.

Prototype

Try PhoenixRise

What I Learned

Understand why students chose an answer

Students had good reasons why answers I considered less effective still worked for them. I learned to acknowledge their reasoning before explaining what could work better.

Use research to guide decisions

Academic articles helped me feel more confident about how my designs could support students’ needs. I learned to use evidence to explain why I included each feature and how it could help.

Keep gamification simple and meaningful

I learned to question whether each game element supported learning or distracted from it. Therefore, I chose a few elements to encourage practice without adding pressure.

Future Steps

Refine AI feedback through testing

Test multiple-choice and AI feedback with more students. This will show whether students understand the feedback and can apply it.

Test their learning

Add a quiz with new academic challenges. This will show how students apply what they learned without hints or guidance.

Expand academic scenarios

Use AI to draft more scenarios and review their relevance. These will give students more academic challenges to practice.

References

List of references

[1] Madigan, D. J., Kim, L. E., & Glandorf, H. L. (2024). Interventions to reduce burnout in students: A systematic review and meta-analysis. European Journal of Psychology of Education, 39, 931–957. https://doi.org/10.1007/s10212-023-00731-3

[2] Cheng, C., & Ebrahimi, O. V. (2023). A meta-analytic review of gamified interventions in mental health enhancement. Computers in Human Behavior, 141, 107621. https://doi.org/10.1016/j.chb.2022.107621

[3] Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191

[4] Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: An experimental investigation of gratitude and subjective well-being in daily life. Journal of Personality and Social Psychology, 84(2), 377–389. https://doi.org/10.1037/0022-3514.84.2.377

[5] Huynh, E., Nyhout, A., Ganea, P., & Chevalier, F. (2021). Designing narrative-focused role-playing games for visualization literacy in young children. IEEE Transactions on Visualization and Computer Graphics, 27(2), 924–934. https://doi.org/10.1109/TVCG.2020.3030464

[6] Butler, A. C., & Roediger, H. L., III. (2007). Testing improves long-term retention in a simulated classroom setting. https://doi.org/10.1080/09541440701326097

[7] Harkin, B., et al. (2016). Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin, 142(2), 198–229. https://doi.org/10.1037/bul0000025

[8] 300+ handmade Doodled Icons for free. (n.d.). Figma Community. Figma. Retrieved October 9, 2026, from https://www.figma.com/design/SGDc5af/IGRJ4w80frsKD/300--handmade-Doodled-Icons-for-free--Community-?node-id=1-75&t=QuLRG5lwtqSn1Fbl-1

More Case Studies

UX Was Here Redesign

A premium space for designers to connect and grow

User Surveys
Onboarding Design
Gamification
ProtoPie

UX Was Here community platform displayed on two laptops

Quebec Winter Carnival Development

Explore the winter carnival through an interactive map

Interactive Map
JavaScript
Microinteractions

Quebec Winter Carnival map inside a snow globe