Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

EduTech Portal: An AI-Powered Student

Assistant Chatbot
Ms. Smita S. Bhosale1; Tejas Kothawade2; Viraj Kharade3;
Adinath Khose4; Suraj Mitake5
1
Assistant Professor, Department of Computer Engineering, Sinhgad Institute of
Technology and Science, Pune
2,3,4,5
B.E. Students, Department of Computer Engineering, Sinhgad Institute of
Technology and Science, Pune

Publication Date: 2025/05/14

Abstract: The increasing number of queries concerning admissions in engineering and polytechnic institutions under the
Department of Technical Education has highlighted the limitations of traditional response methods. To ensure smoother
communication and timely dissemination of information, deploying a centralized AI-powered chatbot emerges as an
efficient and scalable solution. Functioning as a virtual assistant, this intelligent system can autonomously manage
frequently asked questions about eligibility, available programs, fee details, scholarship opportunities, hostel facilities, and
placement timelines. Operating 24/7, the chatbot alleviates the pressure on administrative staff, improves user satisfaction,
and provides valuable data for continuous service optimization. Through machine learning, it evolves by learning from
interactions, enhancing the precision and contextual relevance of its responses. The system not only expedites query
resolution but also reduces the dependency on manual efforts for routine tasks. Additionally, its integration within the
academic framework ensures consistent and fair information delivery. With built-in Natural Language Processing (NLP)
capabilities, the chatbot supports multilingual communication, expanding access and enabling students and their families
to make well-informed decisions.

Keywords: AI-Powered Chatbot, Virtual Assistant for Education, Admission Support System, Automated Query Handling, NLP,
Machine Learning, 24/7 Student Support, Centralized Communication, Digital Education Assistant, Smart Education Technology.

How to Cite: Ms. Smita S. Bhosale; Tejas Kothawade; Viraj Kharade; Adinath Khose; Suraj Mitake (2025). EduTech Portal:
An AI-Powered Student Assistant Chatbot. International Journal of Innovative Science and
Research Technology, 10(5), 122-133. https://doi.org/10.38124/ijisrt/25may390

I. INTRODUCTION AI-powered chatbots presents a viable and innovative

solution. These intelligent systems act as virtual assistants,
The admission process for engineering and polytechnic capable of delivering instant, accurate, and automated
institutions governed by the Department of Technical responses to frequently asked questions around the clock. By
Education is a pivotal phase for students, parents, and leveraging cutting-edge technologies such as Artificial
educational stakeholders. This period typically experiences a Intelligence (AI), Machine Learning (ML), and Natural
significant influx of inquiries related to admission Language Processing (NLP), the proposed chatbot system
procedures, eligibility requirements, tuition fees, offers a scalable, user-friendly, and accessible platform that
scholarships, hostel facilities, course offerings, and transforms how information is disseminated in academic
placement opportunities. Traditionally, these queries are environments. This integration not only reduces the burden
managed through conventional methods such as telephone on institutional staff but also ensures consistent and equitable
conversations, email communication, or in-person visits to information access, thereby empowering stakeholders to
campuses. However, these approaches are often time- make informed decisions during the admission process.
consuming, inefficient, and inconsistent, leading to delays
and a poor user experience. Educational institutions II. PROBLEM STATEMENT
frequently encounter difficulties in coping with the high
volume of queries, resulting in the diversion of valuable During admission seasons, engineering and polytechnic
administrative resources toward handling repetitive institutes experience a sharp increase in queries related to
questions. Consequently, both institutions and stakeholders eligibility criteria, fee structures, placements, and other
face challenges—universities struggle with operational concerns. Traditional methods such as phone calls, emails, or
inefficiencies, while students and guardians are left seeking face-to-face interactions consume significant time and effort
timely and reliable information. To mitigate these issues and for both administrative staff and prospective applicants.
enhance the efficiency of communication, the adoption of Implementing an AI-powered chatbot offers an effective

IJISRT25MAY390 www.ijisrt.com 122

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
solution by delivering round-the-clock support, automating downloadable PDF reports of recommended colleges for
responses to frequently asked questions, easing the burden on offline review and informed decision-making.
staff, and ensuring that students and their families can
conveniently access accurate information. A. Key Features:

III. PROPOSED SYSTEM  Automated Chatbot: Instantly resolves admission-related

queries without human intervention.
The proposed system introduces a centralized AI-  24/7 Availability: Offers continuous access, including
based chatbot designed to streamline admission-related outside working hours.
queries for engineering and polytechnic institutes under the  Multilingual Support: Supports English, Hindi, and other
Department of Technical Education. It delivers real-time, regional languages.
accurate responses to frequently asked questions about  AI & ML Integration: Uses AI for smart responses and
eligibility, admission timelines, fees, scholarships, hostel ML to learn from user behavior.
availability, course options, and placements.  User Authentication: Secure Sign In/Sign Up for
personalized services.
Integrated into the official admission portal, the system  College Recommendation Engine: Recommends colleges
ensures 24/7 accessibility. It employs Artificial Intelligence based on scores, preferences, and other criteria.
(AI) for smart replies, Machine Learning (ML) to adapt  PDF Report Generation: Allows users to download
through user interactions, and Natural Language Processing customized college recommendation reports.
(NLP) to handle multilingual input effectively.  Interactive Interface: User-friendly design for smooth
engagement.
The platform includes secure Sign In/Sign Up features,  Data Insights: Gathers analytics to support institutional
allowing personalized access and data privacy. A key decisions.
addition is the College Recommendation System, which  This system reduces manual effort, improves user
uses user inputs—like academic scores and preferences—to experience, and provides institutions with valuable
suggest suitable colleges. Users can also generate insights for better planning.

IV. SYSTEM ARCHITECHTURE

Fig 1 System Architechture of AI-Powered Student Assistant Chatbot

 User Interaction:  The input is analyzed by the Natural Language

Understanding (NLU) toolkit, which includes:
 The process starts when a user submits a query through
the chatbot interface.  Intent Classifier: Determines the purpose of the query.

 NLU Toolkit Processing:  Entity Extractor:

IJISRT25MAY390 www.ijisrt.com 123

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
 Identifies key information such as names, dates, or  Model Capabilities:
categories. BERT captures bidirectional context in text, allowing
it to understand complex queries by considering both the
 Intent and Entity Flow: previous and following words in a sentence. It provides high
 The detected intents and entities are sent to the Dialog precision in understanding the semantic meaning of queries.
Management Tools for decision-making.
 Impact on the System:
 Dialog Management System: BERT enhances the chatbot’s ability to handle
ambiguous and complex queries. It ensures accurate
 Chooses suitable actions based on the core model and responses by leveraging deep context, improving overall
past interactions. user satisfaction and engagement.

 Uses predefined logic and learning models to generate B. K-Nearest Neighbors (KNN) for Query Suggestion :
relevant responses.
 Steps:
 Data Integration:
 Extract feature vectors from incoming user queries.
 Pulls necessary data from APIs and Databases to enrich  Store feature vectors and corresponding responses from
the response. past queries.
 Ensures dynamic, real-time information like admission  Calculate the similarity between the new query and
deadlines, fee updates, or placement stats. stored vectors.
 Retrieve the K most similar queries based on similarity
 FAQ Retrieval System: calculation.
 Return the best matching response from the retrieved
 For straightforward queries, the dialog system accesses a queries.
Set of FAQs to provide quick, relevant answers.
 Explanation:
 Document Search Module: KNN is a memory-based algorithm that identifies the
most similar past queries to suggest responses. By
 For complex queries, the system uses a search module to comparing feature vectors, it retrieves the most contextually
retrieve information from structured Documents or similar queries and their responses, ensuring relevant
Webpages. suggestions for the user.
 Response Generation and Delivery:  Model Capabilities:
KNN is a simple, instance-based learning algorithm
 The system formulates the final response and delivers it that identifies similar queries by calculating the distance
back to the user via the chat interface. between feature vectors. It is flexible and requires minimal
training, making it suitable for real-time query suggestions.
V. PROPOSED ALGORITHMS
 Impact on the System:
A. BERT for Natural Language Processing (NLP) : KNN improves response efficiency and accuracy by
suggesting relevant past queries, reducing ambiguity and
 Steps: speeding up response time. It enhances the chatbot’s ability
to provide quick, relevant suggestions based on historical
 Preprocess the input query (tokenization, cleaning). data.
 Fine-tune BERT on domain-specific datasets related to
college admissions. C. Text-to-Speech (TTS) for Voice Responses :
 Feed the query into the BERT model to generate
contextual embeddings.  Steps:
 Use embeddings to detect user intent and extract relevant
information.  Generate a text-based response to the user query.
 Generate a context-aware response based on detected  Convert the text response into phonetic units.
intent.  Synthesize the speech from the phonetic units using a
TTS engine.
 Explanation:  Output the synthesized speech as an audio response.
BERT is a powerful NLP model that captures deep
contextual relationships within a sentence by considering  Explanation:
both preceding and succeeding words. Fine-tuning it on The TTS system converts the text-based response into
admission-specific data allows it to understand complex, spoken language, enhancing accessibility and interactivity.
domain-specific queries with high accuracy. By using natural-sounding speech synthesis, it ensures that

IJISRT25MAY390 www.ijisrt.com 124

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
users can interact with the chatbot in a more intuitive and VI. COMPARISON WITH EXISTING SYSTEMS
engaging manner.
AI-powered chatbots outperform traditional methods
 Model Capabilities: by offering superior availability, faster response times, and
TTS systems can generate natural-sounding speech by enhanced user convenience, supported by multilingual
converting text into audio, supporting a range of languages capabilities, voice interaction, and advanced data analytics.
and accents. Advanced TTS models focus on realistic Unlike static conventional systems, AI chatbots leverage
pronunciation, tone, and prosody to improve the user machine learning and natural language processing to
experience. continuously improve accuracy and adaptability while
enabling real-time analytics and personalized interactions.
 Impact on the System: These features ensure seamless scalability, reduced
TTS improves accessibility and user engagement, manpower requirements, and a more intuitive UI with
offering voice-based interaction that is especially beneficial intelligent query suggestions and centralized data
for users who prefer auditory responses or have visual management.
impairments. It makes the chatbot more interactive and
inclusive.

Fig 2 Traditional Methods vs AI-Powered Chatbot

In addition, these chatbots are capable of handling VII. MATHEMATICAL MODEL

large volumes of simultaneous queries, making them ideal
for high-demand environments like educational institutions The AI-powered Student Assistant Chatbot for
during admission seasons. Their ability to learn from engineering and polytechnic institutions is modeled as a
interactions ensures they evolve with user expectations, multi-component system that delivers intelligent, real-time
providing increasingly accurate and context-aware support for admission-related queries. The system is defined
responses. Through centralized dashboards, administrators as a tuple representing its various functional modules, each
can monitor trends, identify common issues, and addressing a specific aspect of the student assistance
continuously improve services. The integration of smart process. These components work collaboratively to ensure
filters, predictive search, and instant document delivery seamless authentication, personalized recommendations,
further simplifies the user journey. Overall, AI chatbots scholarship guidance, and responsive query handling. By
represent a shift toward intelligent automation that leveraging natural language understanding and structured
empowers both users and institutions alike. data processing, the model ensures both accuracy and
accessibility for diverse user needs. Each module is
designed to operate autonomously while contributing to the
system’s overall decision-making process. The modular
architecture also facilitates future integration of additional
educational services, making the platform extensible and
adaptable to evolving institutional requirements.

IJISRT25MAY390 www.ijisrt.com 125

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
 The System is Defined as a Tuple:

IJISRT25MAY390 www.ijisrt.com 126

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

IJISRT25MAY390 www.ijisrt.com 127

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

IJISRT25MAY390 www.ijisrt.com 128

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
VIII. RESULTS

The AI-powered student assistant chatbot website boasts an intuitive and user-centric interface, ensuring seamless navigation
through essential features such as secure login, a customized dashboard, and an up-to-date college listing. Additionally, it includes
a tailored recommendation system, comprehensive scholarship details, and a responsive chatbot to offer real-time assistance
throughout the admission journey. Designed to streamline the decision-making process, the platform leverages natural language
processing and data-driven insights to enhance accessibility and student engagement.

Fig 3 Login Page

Fig 4 Dash Board

IJISRT25MAY390 www.ijisrt.com 129

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

Fig 5 Exploring Colleges

Fig 6 Working of College Recommendation

IJISRT25MAY390 www.ijisrt.com 130

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

Fig 7 Scholarships

Fig 8 Contact

Fig 9 FAQ

IJISRT25MAY390 www.ijisrt.com 131

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390

Fig 9 Working of Chatbot

Fig 10 Chatbot Preview

IJISRT25MAY390 www.ijisrt.com 132

Volume 10, Issue 5, May – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25may390
IX. CONCLUSION [6]. Ahmed Elragal, Ali Ismail Awad, Ingemar
Andersson, “A Conversational AI Bot for Efficient
The AI-based student assistant chatbot provides a Learning: A Prototypical Design,” ResearchGate,
unified platform for addressing admission-related queries in 2023. https://www.researchgate.net/publication/3847
engineering and polytechnic institutes. It automates 94442_A_Conversational_AI_Bot_for_
responses to FAQs on admissions, fees, scholarships, and Efficient_Learning_A_Prototypical_Design
placements, making information more accessible and [7]. Statista Infographics. (Jun. 16, 2023). Infographic:
reducing staff workload. With support for multiple Chat GPT Sprints To One Million Users. [Online].
languages and devices, its NLP-driven interface ensures Available: https://www.statista.com/chart/29174/tim
smooth, user-friendly communication. Additionally, it offers e-to-one-million-users
valuable data insights to the Department of Technical [8]. D. R. Raban and A. Gordon, “The evolution of data
Education for ongoing service improvement. science and big data research: A bibliometric
analysis,” Scientometrics, vol. 122, no. 3, pp. 1563–
FUTURE WORK 1581, Mar. 2020, doi: 10.1007/s11192-020-03371-2.
[9]. C. Bartneck, “The end of the beginning: A reflection
The AI-based student assistant chatbot can be on the first five years of the HRI conference,”
expanded to offer academic support like exam updates, Scientometrics, vol. 86, no. 2, pp. 487–504, Feb.
internships, and career guidance. Real-time integration with 2011, doi: 10.1007/s11192-010-0281-x
institutional systems enables alerts on seat availability, [10]. E. Magnone, “The extreme case of terrorism: A
deadlines, and admission status. Voice and multilingual scientometric analysis,” Scientometrics, vol. 101, no.
support increase accessibility for diverse users. Personalized 1, pp. 179–201, Oct. 2014, doi: 10.1007/s11192-014-
responses based on student profiles improve relevance. It 1378-4
can also predict trends, manage post-admission queries, and [11]. [11] J. Devlin, M.-W. Chang, K. Lee, and K.
engage alumni through events or mentorship. Multi-platform Toutanova, “BERT: Pre-training of deep bidirectional
deployment, including apps and messaging services, ensures transformers for language understanding,” 2018,
broader reach and ease of use. arXiv:1810.04805. https://arxiv.org/abs/1810.04805

REFERENCES

[1]. Franklin Parrales-Bravo, “A Chatbot Solution to

Manage Student Questions About Payments and
Enrollment in University,” ResearchGate, 2024.
https://www.researchgate.net/publication/384677433
_CSM_A_Chat
bot_Solution_to_Manage_Student_Questions_About
_Payments_and_Enrollment_in_University
[2]. Mohammad Amin Kuhail, Hadeel H. Al Khateeb,
Sarah M. Alghamdi, Khaled Shaalan, “Assessing the
Impact of Chatbot-Human Personality Congruence on
User Behavior: A Chatbot-Based Advising System
Case,”ResearchGate,2024. https://www.researchgate
.net/publication/380749206_Assessing_the_Impact_o
f_Chatbot-Human_Personality_Congruence_on_
User_Behavior_A_Chatbot-based_Advising_
System_Case
[3]. Yun-Cheng Wang, Jintang Xue, Chengwei Wei, “An
Overview on Generative AI at Scale With Edge–
CloudComputing,”arXiv,2023.
https://arxiv.org/abs/2306.17170
[4]. [4] Omar Mubin, Fady Alnajjar, Zouheir Trabelsi,
Luqman Ali, “Tracking ChatGPT Research: Insights
from the Literature and the Web,” ResearchGate,
2024. https://www.researchgate.net/publication/377
617512_Tracking_ChatGPT_Researc
h_Insights_from_the_literature_and_the_web
[5]. Lamya Benaddi, Charaf Ouaddi, Abdeslam Jakimi,
Brahim Ouchao, “A Systematic Review of Chatbots,”
IEEEXplore,2024. https://ieeexplore.ieee.org/docum
ent/10542991

IJISRT25MAY390 www.ijisrt.com 133