Section A: Project Information
VisionARi is a bilingual, cloud-based platform tailored to Hong Kong’s new Primary Science curriculum. It blends voice-first AI Tutor dialogue into short animated videos, mobile augmented-reality (AR) scenes, and summary quizzes, into one continuous experience that follows the 5E inquiry model (Bybee et al., 2006) while providing teachers a dashboard that converts raw interaction data into clear, AI-generated progress reports.
The curriculum change arriving in 2025-26 asks generalist primary school teachers to deliver abstract science topics. Interviews with 50+ teachers reveal three major pain points: (i) most teachers lack a STEM background, (ii) resources are scattered, and (iii) digital resources and EdTech solutions are English-centric. Students, meanwhile, are used to interactive media and disengage from static worksheets.
VisionARi’s approach is grounded in pilots that began at HKUST, where an early AR+AI prep module cut remedial lab time by around 80% and improved knowledge retention by around 55%. The pilot has since expanded across four universities, a vocational institute and three secondary/primary schools, generating positive feedback and upcoming presentations at ICLT 2025 and CPLAS 2025. This record shows the architecture works across subjects, age groups and teaching styles.
The platform runs on ordinary iPads/tablets and mid-range phones already ubiquitous in the Hong Kong education landscape. Content will roll out in phases, starting with P4 units in 2025-26, P5 in 2026-27, and P6 in 2027-28 to mirror the Education Bureau (EDB) timeline. Over-the-air updates keep assets aligned with future syllabus changes. We expect to alleviate the anxiety and uncertainty amongst primary science teachers related to the Primary Science roll-out through multimodal and ready lesson plans.
For students, VisionARi’s potential impact is beyond novelty. The AI Tutor adapts its prompts to each learner’s pace, giving hesitant speakers a private space to practice explanations, while the continuity from P4 through P6 builds a coherent mental model of science concepts. Because the same architecture will later extend to bridge activities for P1–P3 and exploratory units for F1–F3, students experience a smooth progression without switching platforms, and teachers gain a stable, ever-growing library that evolves with them.
Section B: Participant Information
| Title | First Name | Last Name | Organisation/Institution | Faculty/Department/Unit | Phone Number | Contact Person / Team Leader | |
|---|---|---|---|---|---|---|---|
| Dr. | Ryu | Fattah | VisionARi Limited | - | rfattah@visionari.app | 66719341 | |
| Mr. | Jungjin | Park | HKUST / VisionARi Limited | MAE / - | jparkak@connect.ust.hk | 59882231 |
|
| Mr. | Jun | Hur | HKUST / VisionARi Limited | MAE / - | jhur@connect.ust.hk | 55790054 | |
| Ms. | Chu Yan | Ho | CityU / VisionARi Limited | Department of Linguistics and Translation / - | chuyanho2-c@my.cityu.edu.hk | 96343844 |
Section C: Project Details
VisionARi began with a stubborn gap observed in HKUST’s undergraduate aerodynamics laboratory. Out of the 60-minute laboratory session, the first 30 minutes were spent reteaching fundamentals that were otherwise meant to be covered through the lab manual, leading to compressed time for data-collection and higher-level discussions. A small pilot that brought an AR+AI intervention eliminated this wasted time by around 80% and saw improved student knowledge retention by around 55%. This pilot has since been extended to a joint-university project with HKUST, PolyU, HKU, and CityU across multiple disciplines (featured in The Chronicle of Higher Education).
The new Primary Science curriculum, starting with P1 and P4 levels in 2025-26, is Hong Kong’s ambitious step towards a strengthened and future-oriented science education. However, interviews with 50+ teachers showed high anxiety about the roll-out of the Primary Science curriculum for a number of reasons: (1) Most primary science teachers do not have a STEM-major background; (2) Resources are scattered across multiple domains and often not aligned with curriculum learning objectives; and (3) There is a lack of Cantonese digital materials when most primary students in Hong Kong communicate best in spoken Cantonese over English or written Chinese.
For these reasons, many schools expressed strong intent to adopt AI technology for teaching. However, almost all educational AI platforms are focused around text-only interactions, usually in English, with limited cohesion to curriculum learning objectives. This is a major barrier at the primary level in Hong Kong. Our hypothesis is that a multimodal and multilingual platform centered around curriculum-mapped videos, voice-first AI dialogue, and interactive AR mini-labs will alleviate the preparation pain for teachers and simultaneously empower them with better student monitoring capabilities. Our framework has already seen initial success at the tertiary level and secondary level, and we believe it will translate to the primary level with minimal friction.
We synthesise multiple technologies that converged in recent years: cloud computing, AR, TTS & STT, and LLM. These proven technology layers now ship on every classroom tablet/iPad and together, they make our core functionalities possible: (1) a mobile app platform with gamified learning progression; (2) animated short-form videos followed by AI-tutor discussion; (3) interactive AR mini-labs guided by AI-tutor; (4) summary quiz aided by AI-tutor; and (5) classroom dashboard with AI-generated student progress reports.
For the development and launch of the VisionARi Primary Science (HK) platform we have a team of 8 interdisciplinary members, collectively working on software development, content development, R&D, and operations. VisionARi is a Cyberport-incubated EdTech company with additional support from the HKUST Entrepreneurship Center and Microsoft for Startups Founders Hub. The platform is built on frameworks previously piloted at the university, secondary, and primary levels, which minimises technical uncertainty.
Prototypes will be demoed to direct beneficiaries (i.e. primary science teachers) in both Cantonese and English from May-June 2025 to collect feedback and commence design iterations. This close collaboration ensures that the final product aligns with curriculum learning objectives and teacher expectations, and thereby making classroom integration a turnkey solution.
Demand signals are already strong. Previous pilots with four universities, one vocational institute, and three secondary/primary schools produced positive learning experiences that are pending to be shared in the form of local education conference presentations (ICLT2025-EdU; CPLAS2025) and student sharings (LTE2025). Upon the Primary Science (HK) roll-out, the main performance metric we will use to evaluate its effectiveness are related to (i) teacher value (comparisons of self-report prep-time and Teacher Net Promoter Score), (ii) learner engagement (weekly active user percentage and module completion ratio), (iv) learning outcome (self-efficacy delta).
-
VisionARi combines distinct technologies and modes that rarely coexist as a single primary-level resource (animated short-video, mobile AR, bilingual voice-AI, assessment, and real-time analytics) and choreographs them inside the 5E inquiry cycle (Engage, Explore, Explain, Elaborate, Evaluate; Bybee et al., 2006). The thread that ties each stage together is ARi, our mascot AI Tutor, who speaks to every learner in a natural voice and personalises each stage.
Engage/Explain: A one-minute animated clip narrated by ARi flows straight into voice dialogue, asking the learner to restate the key idea.
Explore/Elaborate: Inside an AR mini-lab, ARi gives step-by-step instructions to the learner, mixed with exploration questions and context-sensitive hints.
Evaluate: After a summary quiz, ARi goes over any difficult questions with the learner.
The innovation lies in the real-time data hand-off: dialogue transcripts decide which 3-D manipulation to highlight next; lab interactions inform the next discussion question; quiz, voice and touch data stream to a teacher dashboard that tags misconceptions and auto-generates strength/weakness reports and even suggested interventions. Every script line, 3-D asset, and quiz is mapped to a curriculum learning objective, so a teacher can drop a module straight into a lesson plan or project the same AR scene in Classroom Mode as a live teaching aid.
Gamification is purposeful, not decorative. Learners collect up to three stars for each module they complete and earn trophies based on the number of stars they collect in aggregate. At least one star must be earned to unlock the next module, turning the activities into an achievable game rather than a worksheet chore. Meanwhile, teachers stay in control through the dashboard. They can lock or unlock modules, and set deadlines for completion, making VisionARi versatile for pre-learning, homework, and revision, thus precisely addressing teacher anxieties about Hong Kong’s new Primary Science curriculum.
From a technical scalability point of view, we follow industry standard cloud deployment architectures to ensure secure and stable scaling with expected user number growth. In the first year of deployment, we only launch the P4 content library, followed by P5 in 2026-27, and P6 in 2027-28. This 3-year rollout plan aligns with EDB’s own rollout plan, with the only difference being that our platform will not launch content for P1-P3 in this initial rollout. This gives us a predictable number for the service addressable market, which our technical stack can gradually scale to.
As the platform is entirely digital, over-the-air updates are possible. Content updates help ensure alignment with any changes to the curriculum framework while UI/UX updates help ensure long-term user engagement with evolving UI/UX trends. Feature upgrades and updates (e.g. new AI model) help adapt to evolving user needs. By the same principle, replicating the platform at scale to other international curricula is feasible as much of the assets developed are universally applicable with relatively minor localisation work.
From an environmental sustainability point of view, schools will be able to cut down drastically on physical teaching aids, which are often outdated and only used once throughout the year.
Schools in Hong Kong are rapidly adopting the BYOD program where each student has access to a dedicated tablet/iPad device. Therefore, the new Primary Science curriculum is arriving in classrooms where hardware is no longer a major limiting factor. VisionARi leverages this ubiquity of tablets and smartphones, engineered to run on ordinary devices.
The platform directly eases the uncertainty teachers report about the Primary Science roll-out. Each P4-P6 unit arrives with a Suggested Lesson Plan, allowing one-stop integration into classrooms. Furthermore, to guarantee continuity, we are already drafting P1-P3 bridge activities and F1-F3 extension units with relevant stakeholders to ensure a seamless path between lower to upper primary, and between primary to lower secondary.
ARi, the AI tutor, personalises support at both ends of the performance spectrum. Struggling learners receive scaffolded prompts that repeatedly surface keywords and key ideas; high achievers get deeper “what-if” challenges inside the same AR scene. Similarly, ARi gives voice to students who may otherwise be shy to raise their hand in class and often overlooked.
We propose several metrics to measure social impact: (i) digital equity (low-spec device success rate); (ii) school adoption curve (by district and banding); (iii) teacher adoption (dashboard usage rate); (iv) student engagement (completion rate). Responsiveness to the evolving needs of the community is achieved through active participation in education conferences, expos, and school-based feedback cycles.
Personal Information Collection Statement (PICS):
1. The personal data collected in this form will be used for activity-organizing, record keeping and reporting only. The collected personal data will be purged within 6 years after the event.
2. Please note that it is obligatory to provide the personal data required.
3. Your personal data collected will be kept by the LTTC and will not be transferred to outside parties.
4. You have the right to request access to and correction of information held by us about you. If you wish to access or correct your personal data, please contact our staff at lttc@eduhk.hk.
5. The University’s Privacy Policy Statement can be access at https://www.eduhk.hk/en/privacy-policy.
- I have read and agree to the competition rules and privacy policy.