Persona AI: A Multi-Agent Framework for Contextually Aware and Persona-Driven Human-AI Interaction
Main Article Content
Abstract
This paper presents the conceptualisation, design, and implementation of Persona AI, an advanced multi-persona communication platform powered by Google’s Gemini API. The system addresses the limitations of standard "single-persona" chatbots by offering a library of curated characters—such as teachers, doctors, and farmers—each governed by structured behavioural descriptors and role-specific knowledge domains. Central to the platform is a dual-mode interaction strategy: Zen Mode, optimised for functional minimalism and power-user productivity, and Baat-Chit Mode, designed with conversational scaffolding to assist digitally illiterate and rural users. The methodology follows a nine-phase systematic development cycle, integrating Human-Computer Interaction (HCI) principles with state-of-the-art generative modelling. Evaluation through User Acceptance Testing (UAT) and performance benchmarking indicates that persona-driven interactions significantly enhance user self-efficacy, reduce cognitive load, and foster higher levels of trust compared to generic AI interfaces. The findings suggest that the future of AI adoption in diverse socio-economic contexts like rural India depends heavily on the transition from tool-centric to persona-centric systems
Article Details
Section
COPYRIGHT
Submission of a manuscript implies: that the work described has not been published before, that it is not under consideration for publication elsewhere; that if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the publisher.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work
- The journal allows the author(s) to retain publishing rights without restrictions.
- The journal allows the author(s) to hold the copyright without restrictions.
References
[1] Persona AI Major Project Synopsis and Presentation, Geetanjali Institute of Technical Studies, Nov. 2025.
[2] Ministry of Education, Government of India, National Education Policy 2020, 2020.
[3] A. Singh and R. Patil, “Voice-enabled agriculture and health advisory systems in India,” 2021.
[4] Appstek Corp, “The leap from standalone LLMs to autonomous agents,” 2025.
[5] UX Tigers, “The Death of the User Interface (Generative UI),” 2025.
[6] Grand View Research, “AI Companion Market Size Analysis 2024–2030,” 2024.
[7] M. Ackerman, The Sociotechnical Gap. Cambridge, MA, USA: MIT Press, 2000.
[8] “Persona-based Prompting Has an Effect on Theory-of-Mind (PHAnToM) Reasoning,” arXiv preprint arXiv:2403.02246v3, 2024.
[9] S. Patel and A. Kulkarni, “GenAI for vernacular learning material in rural India,” 2023.
[10] UN News Centre, “775 million people in the world are non-literate,” 2012.
[11] University of Tennessee, “Personalised AI Assistants,” 2024.
[12] MultiResearch Journal, “Rural India Broadband Penetration Statistics,” 2024.
[13] M. Carolan et al., “AI usage surge among US adults 2024–2025,” Menlo Ventures, 2025.
[14] Google AI, “Google Gemini 3.1 Pro Technical Report,” 2025.
[15] “Role-play prompting as an implicit chain-of-thought trigger,” arXiv preprint arXiv:2308.07702v2, 2023.
[16] X. Hu et al., “Theoretical Foundations of Persona Prompting,” 2024.
[17] T. Lutz et al., “Role adoption formats and demographic priming strategies,” ACL Anthology, 2025.
[18] “Name-Based Priming to Reduce Stereotyping,” arXiv preprint arXiv:2507.16076v1, 2025.
[19] Patel, M., Choudhary, N. (2017). Designing an Enhanced Simulation Module for Multimedia Transmission Over Wireless Standards. In: Modi, N., Verma, P., Trivedi, B. (eds) Proceedings of International Conference on Communication and Networks. Advances in Intelligent Systems and Computing, vol 508. Springer, Singapore. https://doi.org/10.1007/978-981-10-2750-5_17
[20] “The expertise asymmetry in prompt engineering,” arXiv preprint arXiv:2503.00681, 2025.
[21] Jagannath University, “Digital divide exacerbated by infrastructure gaps,” 2026.
[22] ICT Works, “HCI for Development (HCI4D) Principles,” 2025.
[23] BMZ FAIR Forward, “Localised AI solutions using vernacular languages,” 2024.
[24] National Health Mission, “Suman Sakhi WhatsApp Chatbot Pilot,” 2024.
[25] Government of India, “Bhashini: National Multilingual AI Platform,” 2022.
[26] “Conversational Scaffolding and Pre-made Prompts,” arXiv preprint arXiv:2602.04109v1, 2026.
[27] “Minimalist Design and Functional Minimalism,” Medium, 2025.
[28] W3C COGA, “Guided Scaffolding for Beginners,” 2024.
[29] ResearchGate, “The AI Assistance Dilemma in Cognitive Tasks,” 2025.
[30] “Improving Self-Efficacy Through Guided Interfaces,” arXiv preprint arXiv:2501.12001v1, 2025.
[31] React.js Documentation, “Scalable Web Architecture for AI Interaction,” 2024.
[32] Google Developers, “Thought Signatures for State Management,” 2026.
[33] Google Cloud, “Long-context capabilities of Gemini models,” 2025.
[34] Google DeepMind, “Interactions API for Stateful Workflows,” 2025.
[35] Letta, “Building Stateful AI Agents with Advanced Memory,” 2025.
[36] Google AI Studio, “Encrypted Thought Signatures in Gemini 3,” 2026.
[37] Supabase, “Supabase for Authentication and Data Isolation,” 2024.
[38] “System prompts precedence over user inputs,” arXiv preprint arXiv:2505.21091v2, 2025.
[39] “Mathematical Foundations of Personalised Generation,” arXiv preprint arXiv:2502.11528v2, 2025.
[40] MDPI, “Intelligent Tutoring and Structured Support,” 2025.