NexGenStudy: An Adaptive, Immersive, and Blockchain- Secured Mobile Learning Gaming Platform Integrating IVR, Unity-Based Gami?cation and Augmented Reality for Next-Generation Education

NexGenStudy: An Adaptive, Immersive, and Blockchain- Secured Mobile Learning Gaming Platform Integrating IVR, Unity-Based Gami?cation and Augmented Reality for Next-Generation Education

Manasa Thouta

Dept. of Information Technology Gokaraju Rangaraju Institute of Engineering and Technology Hyderabad, India

Kalluri Sai Spoorthy

Dept. of Information Technology Gokaraju Rangaraju Institute of Engineering and Technology, Hyderabad, India

Eluri Jaya Sri

Dept. of Information Technology Gokaraju Rangaraju Institute of Engineering and Technology Hyderabad, India

Chilukuri Sujatha

Assistant Professor, Dept. of Information Technology Gokaraju Rangaraju Institute of Engineering and Technology Hyderabad, India

Abstract – Modern digital education platforms have mainly followed the static, one-size-ts-all modes of delivery that are not able to support the cognitive and motivational needs of the present-day learner. NexGenStudy, a unique Flutter-based mobile learning app that integrates adaptive computing on AI, real-time cloud computing, voice-based AI using Interactive Voice Response (IVR), Unity-based gamication, Augmented Reality (AR)-based simulation, and blockchain-authenticated academic credential management all into one unied ecosystem and smart learning platform, is introduced in this paper. The platform uses machine learning and algorithmically adjusts study plans and assessment difculty to the performance trajectory of a particular learner, and an AI chatbot, articial intelligence driven by the Gemini API provides tutoring context-awareness and real-time doubling from doubts. An original IVR module enhances the accessibility of learners who have low-connectivity conditions by the use of voice-based ofine doubt solving – a non-existing feature in the literature. Unity 3D game makes immersive, level-based learning quests that convert academic assignments into goal-based challenges, which dramatically increases engagement and retention. Ethereum, through Ganache and blockchain tech-nology, can be used to implement an immutable, nondestructible academic progress records ledger, which allows portability of academic credentials across all institutions and ensures privacy. Firebase is the scalable real-time cloud backend that is used to host authentication and personalized content delivery (on Fire-store) as well as analytics about a community. The experimental objectives project 50% higher conceptual understanding, 50%higher peer collaboration, and 100% data privacy through decen-tralized record-keeping. NexGenStudy is a radical amalgamation of technologies, which, alone, can be found in previous literature but not at this scale in one adaptive learning platform.

Index Termsadaptive computing, voice-based AI interaction, AR-

based simulation, real-time cloud processing, AI computa-tion, blockchain, gamication, IVR, Unity 3D, authentication, articial intelligence, ofine doubt solving, augmented reality ed-ucation, machine learning, personalized, smart education, digital unied learning ecosystem

1. Introduction

   The digital learning sphere has reached a height we did not imagine possible with the deployment of smartphones,

   cloud and articial intelligence at a dizzying speed but the e- learning platform is unable to utilize the technology in the right direction for learning. Most e-learning delivers linear drag and drop content or a series of activities without analyzing learner performance, behavior, and readiness. [1]. According to recent research, students are becoming less motivated during their learning activities owing to a lack of interactivity. [2].

   Three persistent challenges motivate the design of NexGen- Study: (i) cognitive disengagement- passive consumption of video-text content without immersive interaction; (ii) support inaccessibility- academic assistance conned to synchronous classroom hours, leaving learners stranded during self-study; and (iii) credential fragmentation- academic records stored in siloed, mutable institutional databases that are neither portable nor privacy-preserving.

   The NexGenStudy is dealing with these challenges by pro- ducing an unprecedented synthesis of 6 technology pillars: (1) ML-driven adaptive content personalization, (2) Gemini-API- powered AI chatbot tutoring with real-time cloud processing,

   (3) IVR-based ofine voice assistance via Twilio, (4) Unity 3D educational gamication with level-based progression, (5) ARCore/ARKit augmented reality simulation for conceptual visualization, and (6) Ethereum-blockchain-backed immutable academic record management. No previously published plat- form combines IVR, Unity gamication, and blockchain into a single mobile learning system. This is the main innovation of this work.

   The remainder of this paper is organized as follows. Sec-tion

   II surveys related work. Section III describes system architecture and design. Section IV details implementation of each module with technical specics and design artifacts. Section V presents results through functional validation of all implemented screens and a consolidated evaluation summary. Section VI concludes the paper.

2. Literature Survey

   1. Conversational AI and Chatbots in Education

      Otermans et al. [1] conducted a review of six recent studies on using conversational AI in higher education. They highlighted its potential for personalized feedback, scalable instruction, and knowledge sharing between institutions. How- ever, the review points out serious limitations. These include errors in AI-generated content, lack of emotional support, and unresolved ethical and policy issues about student aware- ness of AI use. NexGenStudy reduces the risk of content inaccuracies by basing the AI tutor on structured Firestore question histories. This approach allows for context-aware and curriculum-focused responses.

   2. AI-Driven Adaptive Learning Pathways

      The study by Abrar et al. [2] compared an AI-driven person- alized learning system to a quasi-experimental six-week study involving 200 undergraduate students. Academic performance was improved by 25%, task completion rate was also better by 25%, and engagement was also better by 15% in the AI-based cohort than in traditional instruction. Scalability limitations and the necessity of privacy-sensitive frameworks as a gap that NexGenStudy directly seals through Firebase-based scalable clouds and Ethereum blockchain-based credential storage are also mentioned in the study.

   3. AI-Generated Quiz Systems

      Wagh et al. [3] presented Quizify, an Android application leveraging the Google Gemini API, Kotlin, and Firebase with an MVVM architecture to dynamically generate curriculum- aligned assessments. This system is said to have cut the workload of teachers by 70% and enhanced the customization of quizzes. NexGenStudy builds on this paradigm, but not only pairs AI-created quizzes with dynamically increasing difculty based on a learners current prociency level (beginner, inter- mediate, or advanced) in Firestore but can also provide fully longitudinal assessment adaptation.

   4. Augmented Reality in Education

      Hincapie´ et al. [4] conducted a bibliometric analysis of 215 AR-in-education publications spanning 20032018, con- rming that AR interventions consistently support kinesthetic and immersive learning, improve motivation and memory retention, and are particularly effective in STEM contexts. The review, nevertheless, determnes the risks of cognitive overload and usability issues that can be attributed to the lack of standardized AR pedagogical design frameworks. NexGenStudy uses AR using ARCore (Android) and ARKit (iOS) with a scaffold to interaction in order to alleviate the effects of cognitive overload.

   5. AI-Based Mobile Learning Recommendation

      Zhu et al. [5] performed a eld study of an AI deep-learning recommendation system of mobile learning on a sample of 400 middle-school students and reported an 85% weekly active use and up to 83% learner satisfaction. The article proves

      the practicability of large-scale content recommendation based on ML and admits the deciencies in the transparency of algorithms and their predictive quality. The ML recommen- dation engine of NexGenStudy is designed to use ne-grained prociency data, which is stored in Firestore and can be used to create subject-specic adaptive playlists of videos, better personalized, which are enhanced with more powerful personalization cues.

   6. Interactive Voice Response in Education

      Shanbag et al. [6] examined the IVR application in learning institutions and reported its benets of 24/7 access, cost- efciency, and scalability in the delivery of information in resource-limited settings. The mentioned limitations are identi- ed to be menu complexity and language/hardware constraints. The reviewed literature does not contain any instances of IVR integration as a supplement to mobile learning, which NexGenStudy provides as it is the rst platform to offer IVR (through Twilio) as an additional ofine system of doubt- solving to reach learners with limited connectivity or mobility.

   7. Gamied Mobile Learning

      Kao et al. [7] applied a BOPPPS-model gamied learning system in a quasi-experimental study on accounting education, reporting signicant improvements in academic achievement and engagement. AlAli [8]discovered signicant academic and attitudinal improvement in 81 gifted students who applied gamied mobile apps.Hardware dependency and the possibil- ity of distraction are mentioned in both papers. The Unity 3D gamication module of NexGenStudy is designed to ensure that game-level advancement is highly structured to follow the milestones of the curriculum so that the entertainment mechanisms do not distract but support academic goals.

   8. Research Gap and Novelty

   Synthesizing the above literature, a clear research gap emerges: no existing platform simultaneously integrates IVR- based ofine voice assistance, Unity 3D academic gamica- tion, and Ethereum blockchain credential management within a single adaptive mobile learning ecosystem. Each technology appears in isolation or in limited two-way combinations. This gap is lled by NexGenStudy, which provides an immersive, comprehensive, and privacy-preserving educational architec- ture

3. System Architecture

   NexGenStudy is structured around a four-layer architecture as illustrated in Fig. 1: (i) the Student Interaction Layer, where the learner interfaces with the application; (ii) the Flutter Mobile Application Layer, encompassing the UI Layer and ve functional modules; (iii) the Backend Services Layer, pro- viding AI computation, cloud data storage, and authentication; and (iv) the Blockchain Layer, ensuring immutable academic credential management.

   At the highest level, the student communicates only through the Flutter-rendered UI Layer, which acts as the single

   point of entry for ve parallel functional modules: (1) IVR Ofine Calling, an ofine doubt resolution voice channel powered by Twilio; (2) AI Tutor Module, a conversational tutor supported by the Gemini API that provides real-time explanations; (3) ML Personalization Engine, a machine learning component that curates adaptive study plans and predictive reminders; (4) Unity Game Module, an embed- ded Unity 3D gamication runtime that communicates with Flutter via a bidirectional message-passing interface via the flutter_unity_widget package; and (5) AR Learning Lab, an augmented reality visualization environment powered by ARCore/ARKit.

   The Backend Services Layer, which consists of three parts, is interfaced with by all ve modules: the Gemini API for AI computation and quiz creation, the Firestore Database for real-time cloud storage of learner proles, session histories, quiz results, and community analytics, and Firebase Auth for secure user authentication. The Firestore Database is further connected to the Blockchain Layer (Ethereum + Ganache), where Web3Dart is used to write veried academic milestones as immutable smart contract events. By ensuring that only veried, server-conrmed accomplishments are permanently recorded on the blockchain, this unidirectional data ow from Firestore to the blockchain guards against client-side manipulation.

   TABLE I NexGenStudy Technology Stack

   Component	Technology / Library
   Mobile framework	Flutter 3.x (Dart)
   AI / NLP engine	Google Gemini 1.5 Flash API
   Cloud backend	Firebase (Firestore, Auth, Functions)
   Gamication engine	Unity 3D (embedded via flutter_unity_widget)
   AR runtime	ARCore (Android) / ARKit (iOS)
   IVR / Voice channel	Twilio Programmable Voice
   Blockchain	Ethereum + Ganache (Web3Dart)
   Smart contracts	Solidity v0.8.x

   1. ML-Based Adaptive Personalization Engine

      Fig. 1. NexGenStudy four-layer system architecture: Student Flutter UI Layer ve functional modules (IVR Ofine Calling, AI Tutor, ML Personalization Engine, Unity Game Module, AR Learning Lab) Backend Services (Gemini API, Firestore, Firebase Auth) Blockchain (Ethereum + Ganache).

4. Implementation Details

   A. Technology Stack and Development Environment

   Flutter (Dart) is the cross-platform mobile framework used in the development of NexGenStudy, which targets iOS 13+ and Android API Level 24+. Table I provides a summary of the entire technology stack. Firestore is the NoSQL real- time database that houses learner proles, quiz histories, community channel data, and session analytics, while Firebase is the main cloud backend. Firebase Authentication uses OAutp sign-in ows and email/password to manage secure user identity.

   Fig. 2. ML personalization pipeline: user quiz performance updates pro- ciency tier in Firestore; content-based ltering selects subject playlist; gradient-boosted classier schedules adaptive study reminders.

   The personalization engine creates subject-specic study playlists using a content-based ltering algorithm enhanced with collaborative signals. A three-state Markov process- S b, S i, and S a, representing the beginner, intermediate, and ad- vanced tiers, respectively, is used to model learner prociency. Quiz performance thresholds b = 0.50 and i = 0.75 control state transitions; a learner who scores below b is kept at Sb, between b and i advances to Si, and above i advances to Sa. The learners current state and selected subject domain are used to lter curated video playlists that are retrieved from Firestore. The flutter_local_notifications package, which is integrated with the device calendar, uses a lightweight gradient-boosted classier trained on session log features (time-of-day, session gap, and streak length) to predict the best push-notication timing for study reminders. The personalization pipeline is shown in Fig. 2.

   1. AI Chatbot Tutor with Real-Time Cloud Processing

      The AI Tutor module uses the gemini 1.5 ash API and a conversation history stored n Firestore to provide context- dependent curriculum-constrained academic advice. Each API invocation is structured as a multi-turn prompt that embeds:

      (i) the learners current subject and prociency tier, (ii) the ten most recent question-answer pairs retrieved from Firestore, and (iii) the incoming query. This context window design allows for a sense of topical continuity between sessions without the need for server-side memory. Firebase Cloud Functions acts as a proxy server layer that calls the Gemini endpoint and returns the result to the Flutter client, allowing for a sub-500 ms median round-trip latency under normal LTE conditions. The interface also displays tappable follow- up suggestion chips based on the response from the model to reduce cognitive load for the next question to ask. A gamied scoring system gives users a score for each AI tutoring session, displayed as a live score counter in the UIs header section.

   2. IVR-Based Ofine Voice AI Interaction

      IVR module, enabled through Twilio Programmable Voice, stands out as the most architecturally distinctive innovation of NexGenStudy in comparison with existing literature. Unlike conventional in-app chatbot interfaces, which require constant internet connectivity, NexGenStudy makes the barrier for academic support a mere single interaction: the student merely needs to tap a single button on the application, which in turn initiates a direct call from the Twilio platform to the students registered phone number. The student does not need to navigate, type, or maintain an active data connection; the AI-driven student doubt resolution process reaches the student as a voice call, meeting the student where they are. [6].

   3. Unity 3D Educational Gamication

      The gamication layer is built within Unity 3D and embedded into the Flutter application using the flutter_unity_widget package, enabling bidirectional communication through a JSON-based message-passing interface. The Unity world is rendered as an open low-poly

      urban environment in which the students avatar navigates subject-specic learning zones (e.g., a Science district, Mathematics tower, Humanities museum). Within each zone, academic interactions are presented as in-world challenges: an NPC character poses a multiple-choice question rendered as oating option panels; correct answers trigger environment transitions, unlock new areas, and award in-game currency. Academic tasksquiz completion, AR lab visits, AI tutor sessionsare mapped to progression events that Flutter sends to Unity via postMessage(), updating the players level and achievements in real time. The game design adheres to Self-Determination Theory [9], embedding autonomy (learner-chosen zone navigation), competence (adaptive question difculty), and relatedness (peer leaderboard) to sustain intrinsic motivation.

   4. AR Learning Lab

      The AR Learning Lab uses the ar_flutter_plugin, which supports both Android and iOS using the native SDKs provided by ARCore and ARKit, respectively. Upon opening the AR Learning Lab, the camera enters plane detection mode. After a horizontal plane has been successfully detected, the learner can choose a 3D object from a list of available models (e.g., Human Heart, Tower of Hanoi, molecular structures, circuit diagrams, astronomical objects).

      The chosen object is then instantiated as a world-anchored SceneKit/ARCore node, allowing for scaling and rotating using pinch and swipe gestures. The annotation system uses AR text labels positioned on specic structures of the object, enabling step-by-step guided explanations without requiring the learner to navigate to a separate reference screen.

      The Tower of Hanoi object also includes disc manipulation, allowing the learner to physically interact with the recursive algorithm in a three-dimensional environment. This interaction modality is based on dual coding theory [4], where both cognitive channels are engaged simultaneously.

   5. Blockchain-Based Academic Credential Management

      The NexGenStudy system utilizes private Ethereum blockchain technology for academic credential storage via the Ganache node, which is integrated through Flutter via the web3dart library. A Solidity smart contract is used to expose a recordAchievement(bytes32 studentId, string subject, uint8 score, uint256 timestamp) method. Each milestone, such as quiz completion, module mastery, AR lab completion, and game level unlocking, is recorded through a call to this smart contract via Firebase Cloud Functions (server-side), ensuring that client-side manipulation is not possible. The studentId eld is a SHA-256 hash of the Firebase UID, ensuring pseudonymity. The data is time-stamped and immutable and can be exported in JSON-LD format for veriable credentials. Table II shows key parameters for the system.

      Parameter	Value / Choice
      Blockchain framework Smart contract language Flutter integration library Trigger source Record type Student identier Credential export format	Ethereum (Ganache local node) Solidity v0.8.x Web3Dart Firebase Cloud Functions (server-side) Achievement event log SHA-256 hashed Firebase UID JSON-LD Veriable Credential

      TABLE II Blockchain Design Parameters

      platforms exible content structure. Study plan generation, AI tutoring, community channels, and AR model catalogs all adjust to the subject context chosen by the learner and stored in Firestore.

   6. Community Learning Dashboard

   The Community Dashboard uses a subject-channel model, similar to real-time messaging systems. Each subject, like #HumanAnatomy, #LinearAlgebra, or #AncientRome, is a Firestore subcollection. Messages go out to all subscribers through Firestores onSnapshot() listener, achieving de- livery latency under 200 milliseconds. Community analytics, including quiz pass rates, module engagement frequency, and peer counts, are gathered on the server side using Firebase Cloud Functions. These functions trigger with Firestore write events and provide data to learners as peer comparison nudges, like 80% of peers in your cohort aced this quiz; try it! All analytics are anonymized through differential privacy perturbation (c = 1.0) before aggregation, ensuring that no information about individual learners can be reconstructed from the aggregated statistics.

5. Results and Discussion

   The following subsections present the implemented appli- cation screens alongside functional validation observations for each module. All screenshots were captured on a physical Android device running the production Flutter build.

   Fig. 3. NexGenStudy home screen: personalized greeting, Wordle-style brain teaser, and four feature naviga- tion tiles.

   Fig. 4. Subject selection: eight cross- disciplinary domains as colour-coded tiles feeding context into the ML en- gine, AI tutor, and AR lab.

   1. Application Home Screen and Daily Brain Teaser

      Fig. 3 shows the NexGenStudy home screen. It welcomes the authenticated learner by name and displays four main feature entry points: AR Learning, Ofine Chatbot (IVR), AI Quizzes, and AI Tutor. A prominent Todays Brain Teaser widget offers a daily vocabulary or concept challenge as a Wordle-style ve-letter guessing game. For example, learners might identify a calculus term from its denition. This en- gaging interaction happens right away when the app opens, eliminating the need for navigation. This design approach helps overcome the initial motivation barrier noted in mobile learning adoption research [8]. The bottom navigation ensures easy access to Home, StudyPath, Community, and Prole at all times.

   2. Topic Selection and Subject Navigation

      Fig. 4 displays the subject-selection screen. t offers eight learning areas: Calculus, Linear Algebra, Quantum Physics, General Relativity, Organic Chemistry, Human Anatomy, World War II, and Ancient Rome. Each area appears as a color- coded tile with an icon, making it easy to tell apart STEM and Humanities subjects. This wide range of subjects shows the

   3. ML-Driven Personalized StudyPath

      Fig. 5 presently indicates that the Personalized StudyPath developed by ML to a learner who has chosen an STEM sub- ject. The engine generates a multi-stage learning programme sequenced, i.e., Step 1 (Core Exam Topic Identication), Step 2 (Foundational Video Lecture Series), and Step 3 (Targeted Problem Walkthroughs): each of which contains actionable sub-instructions. The study path is created based on the Gem- ini API conditioned on the topic of the learner, their level of prociency and their previous patterns of sessions maintained in Firestore and regenerated on-demand once performance information shows improvement in the tier. This adaptive curriculum design can be seen as a direct response to the limitation on content delivery in the literature [2] namely offering a dynamically scaffolded pathway of learning as opposed to a pre-written xed syllabus.

   4. AI Tutor Module

      Fig. 6 illustrates a live AI Tutor. The learner types in the query of what is environment and gets a more detailed, speech structured answer that denes the concept, differentiates biotic and abiotic items and the Socratic end of learning question.

      The tutoring session is gamied with a live score counter shown in the top-right header, which allows obtaining points each time of interaction. The conversational interface has been architecturally supported by Gemini 1.5 Flash and the entire Firestore conversation history is injected into the context of the input (as context), allowing uid multi-turn conversation. The session demonstrates the models ability to deliver aca- demically structured responses (denition elaboration example comprehension check) rather than raw information retrieval.

      capability that isnt found in the reviewed mobile learning literature [6].

      F. AR Learning Lab

      Fig. 8 shows the display of the AR Learning Lab that features a high-delity 3D model of a human heart created in ARCore plane detection. It is possible to switch between cata- logued 3D models using the bottom model-selector bar. Heart model It is a completely gesture-interactive model: learners are able to spin 360, zoom to explore the chambers of the heart, and activate annotation overlay, which labels structural features. This embodied spatial interaction is a realisation of the dual-coding pedagogical approach [4], learners encode both the visual form of a three dimensional form and symbolic anatomical terms, and this encoding conceptually encodes far more than is the case with textbook diagrams.

      Fig. 5. ML-generated Personalized StudyPath: three-step sequenced cur- riculum adapted to the learners pro- ciency tier.

      Fig. 6. AI Tutor: Gemini 1.5 Flash delivers Socratic multi-turn dialogue with gamied score tracking in ses- sion header.

   5. IVR Ofine Voice AI Call Interface

   Fig. 7 shows the IVR modules in-app trigger screen, titled AI Voice Call Demo. A single prominent Call AI Assistant button starts an outbound Twilio Programmable Voice call to the learners registered phone number. Unlike standard

   Fig. 7. IVR interface: single-tap Call AI Assistant triggers a Twilio voice session for ofine doubt reso- lution without active internet.

   Fig. 8. AR Learning Lab: gesture- interactive 3D human heart rendered via ARCore with anatomical annota- tion overlays.

   in-app chatbot interfaces that need constant internet access, NexGenStudy makes academic support just a tap away. The AI-powered doubt resolution session comes as an incoming voice call, meeting the student right where they are. Once the call connects, the Twilio TwiML ow presents a voice menu of subject categories. The learners spoken question is transcribed by Twilio ASR, checked against the Firestore FAQ database, and answered using text-to-speech. This works without needing an active internet connection on the mobile data channel. This feature offers a truly unique ofine fallback

   G. Community Learning Dashboard

   Fig. 9 shows the Community Dashboard in use. The left panel has a sidebar organized by subject area: Science, Mathematics, Humanities, and Technology. Active channels include #Human Anatomy, #Organic Chemistry, #Quantum Physics, #Calculus, and #Linear Algebra. The right panel displays a real-time chat stream with peer messages that have timestamps. These messages are delivered through Firestore streaming with a delay of less than 200 ms. The channel design, inspired by Discord, makes it easier for peers to

   collaborate by using familiar interaction patterns in specic academic subjects [5].

   H. Unity 3D Game: In-App Mobile View

   The Unity 3D educational game has been rendered in the Flutter application as illustrated in Fig. 10. The player moves a humanoid robot avatar that moves through a city scene of low- polygamy. A discussion is ongoing between two academics: an in-world NPC asks a question with four multiple-choice option buttons, which is then oated on a dialogue panel. This contextualized type of question-answer mechanic places academic assessment within the spatial discourse of the game world, rendering taking quizzes an outcome of world explo- ration, rather than a formal activity.

   Fig. 11. Unity 3D game world map (top-down editor view): richly detailed low-poly urban environment with subject-specic districts, progressively un- locked through academic milestone completion.

   J. Summary of Functional Validation

   Table III shows a summary of functional validation for all the implemented modules. It compares the observed behavior to the design goals.

   Module	Design Objective	Observed Result
   ML StudyPath	Prociency- adaptive plan	3-step dynamic curriculum generated per subject/tier
   AI Tutor	Socratic, context- aware tutoring	Multi-turn Gemini responses with comprehension checks
   IVR	Ofine voice doubt resolution	End-to-end Twilio call ow; ASR + TTS veried
   AR Lab	Embodied 3D con- cept interaction	Heart & Tower of Hanoi: gesture- interactive, annotatable
   Unity Game	Curriculum- embedded gamication	NPC quiz challenges in open- world navigation
   Community	Real-time peer col- laboration	Sub-200 ms Firestore delivery across subject channels
   Blockchain	Immutable creden- tial storage	Achievement events recorded as Solidity contract logs

   TABLE III Functional Validation Summary

   Fig. 9. Community Dashboard: Discord-style subject channels with real-time Firestore-streamed peer messaging.

   Fig. 10. Unity 3D game on mobile: robot avatar navigates low-poly urban world and engages NPC academic challenges.

   I. Unity 3D Game: World Map (Editor View)

   Fig. 11 shows the top-down view of the entire game world map in the Unity Editor. It is a dense urban grid with residential quarters, a stadium, hospital areas (marked H), road systems, and various buildings of different architecture in the style of low-poly art. All the districts are matched to particular academic subject area, clicking on a building triggers the con- tent of the related subject. The globe is increasingly opening up as students pass through milestones in their academic work, and acts as a spatial manifestation of curriculum advancement.

6. Conclusion

This paper introduces NexGenStudy, a next-generation adaptive revolutionary learning platform integrating AI com- putation, real-time cloud computation, voice-driven AI inter- action with IVR, Unity 3D gamication, AR-based simulation and Ethereum blockchain credential management into one unied Flutter application. The novelty of the platform in its architecture is that the platform has three features that are not found in the previous literature: (1) IVR-based ofine voice-AI tutoring to reach low-connectivity learners, (2) Unity 3D teaching gamedication closely integrated with curriculum- level progression through a two-way Flutter message bus, and (3) Ethereum smart-contract-based storage of academic credentials to allow cross-institutional and tamper-resistant portability.

Through an immersive, privacy-sensitive and holistic de- sign, NexGenStudy permeates the three main drawbacks of modern e-learning: cognitive disengagement, inaccessibility of support, and fragmentation of credentials.

Acknowledgment

We thank Mrs. Chilukuri Sujatha, Assistant Professor, Department of Information Technology, Gokaraju Rangaraju Institute of Engineering and Technology, Hyderabad, for her invaluable guidance and mentorship throughout the develop- ment of NexGenStudy.

References

1. P. C. J. Otermans, S. Demetriadis, and D. Richards, Editorial: The role of conversational AI in higher education, Education and Information Technologies, 2025.

2. M. A. W. Aboraya, R. Abdulghafor, K. P. Subramanian, Y. Al Husaini, and M. Al Hussaini, AI-Powered Learning Pathways: Personalized Learning and Dynamic Assessments, Journal of Educational Technology Systems, 2025.

3. K. Wagh, B. Sahu, A. Rathod, R. Wankhade, and A. Borkar, Quizify: Intelligent Quiz Generator Powered by AI, International Journal of Advanced Computer Science Applications, 2025.

4. M. Hincapie´, C. D´az, A. Valencia, M. Contero, and D. Gu¨emes- Castorena, Educational Applications of Augmented Reality: A Bibliometric Study, Computers and Education, 2021.

5. Y. Zhu, W. Dai, and Q. Kang, The analysis of AI-based mobile learning in students open teaching recommendation system, Mobile Information Systems, 2025.

6. R. Shanbag, P. K, Akshay, and Kanthi, Interactive Voice Response for Educational Institution, International Journal of Engineering Research and Technology, 2023.

7. M.-C. Kao, Y.-H. Yuan, and Y.-X. Wang, The study on designed gamied mobile learning model to assess students learning outcome of accounting education, Educational Technology Research and Development, 2023.

8. R. AlAli, Revolutionizing Education: Assessing the Impact of Mobile Learning Apps on Academic Success and Attitudes, Journal of Educational Computing Research, 2025.

9. S. de Sousa Borges, V. H. S. Durelli, H. M. Reis, and S. Isotani, A Systematic Mapping on Gamication Applied to Education, Proceedings of the ACM Symposium on Applied Computing, 2014.

10. H. Ukgoda, Gamication in Education: Its Impact on Engagement, Motivation, and Learning Outcomes, Education Sciences, 2025.