© JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

Scheme Setu: A Chatbot for Government Schemes

SHREYAS S1, TUSHAR TIWARI2, THEJAS V3, VIKRAM4
1, 2, 3, 4
Department of ISE, Presidency University, Bangalore, India

Abstract- In the ever-changing landscape of digital engagement. At its core, this innovative chatbot
services and government initiatives, our project amalgamates the prowess of advanced Natural
embarks on a mission to empower citizens with a Language Processing (NLP) algorithms and state-of-
revolutionary chatbot known as SchemeSetu. This the-art machine learning models. These technological
intelligent chatbot serves as a central information marvels converge to create an intuitive,
hub, consolidating crucial details on government- conversational interface—a digital ally that
sponsored loans and insurance schemes from empowers individuals to navigate the labyrinthine
various sources. Harnessing cutting-edge realm of government services effortlessly. The
technologies and natural language processing, bedrock of our initiative lies in the assimilation of
Scheme Setu acts as a unified gateway to essential credible data from revered repositories such as the
financial assistance programs. Drawing National Bank for Agriculture and Rural
information from reputable institutions like Development (NABARD) and the Reserve Bank of
NABARD and RBI, our innovation not only India (RBI). This partnership ensures not only the
simplifies access but also enriches the user accuracy and credibility of the information but also
experience. Individuals can effortlessly explore, instills a sense of trust and reliability among users
comprehend, and benefit from a range of seeking financial guidance. The Mega Bot is not just
governmental financial offerings. With Scheme a digital innovation; it's a catalyst for change—a
Setu, our aim is to transform how individuals catalyst that propels financial literacy, fosters
interact with and access government services, inclusivity, and redefines the paradigm of citizen
fostering financial literacy and promoting empowerment. By offering a unified platform for
inclusivity in financial matters. comprehending, accessing, and capitalizing on an
array of governmental and financial offerings, our
I. INTRODUCTION initiative aims to dismantle barriers, democratize
opportunities, and inspire a wave of financial
In an era defined by the relentless evolution of digital empowerment across diverse strata of society. In
services and the rapid proliferation of government essence, our project embodies a transformative
initiatives, our project stands as a beacon of vision—one that transcends the conventional
innovation, poised to empower citizens through a boundaries of citizen-state interactions. Through
groundbreaking solution: Scheme Setu. With an Scheme Setu, we embark on a journey to redefine
unwavering commitment to harnessing technology accessibility, knowledge dissemination, and
for societal betterment, we embark on a mission to inclusivity in government-sponsored financial
revolutionize the accessibility and understanding of services, catalyzing a future where every individual
crucial government-sponsored loans and insurance possesses the tools to navigate and leverage the vast
schemes. Scheme Setu represents an ambitious landscape of governmental support.
endeavor, born from the convergence of cutting-edge
technologies and an acute understanding of the II. OBJECTIVES
citizen's needs. It serves as an intelligent and
interactive gateway—a singular, comprehensive The primary objectives of this research endeavor
reservoir of knowledge sourced from diverse and encompass the establishment of a Comprehensive
reputable channels. Our pursuit transcends mere Information Hub dedicated to government-sponsored
information aggregation; it encapsulates a vision to loans and insurance schemes. This hub will function
democratize access to essential financial assistance as an expansive repository, consolidating diverse and
programs, transforming the landscape of citizen-state vital information from various sources for easy

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 119

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

access and understanding. Emphasizing user Reporting encompass a comprehensive overview of

interaction, an engaging and intuitive interface will the tools, methodologies, challenges, and solutions
be designed to promote seamless exploration, query, applied throughout the development process,
and comprehension of available financial assistance providing a detailed report on the technical aspects,
programs. To ensure credibility, the system will workflow, and outcomes achieved. This holistic
provide reliable and accurate information sourced methodology addresses data collection, model
from esteemed institutions such as NABARD and implementation, integration with external platforms,
RBI, positioning the chatbot as a trusted source for automation for updates, user-centric design, and
individuals seeking financial guidance. Moreover, the establishes a robust framework for ongoing
implementation of a personalized recommendation enhancement and maintenance of the chatbot.
system within the chatbot aims to offer tailored
suggestions based on individual needs and IV. CHUNKED DATA STREAMING WITH
preferences, guiding users towards the most relevant LARGE LANGUAGE MODELS (LLMS)
government initiatives. Furthermore, this initiative
seeks to pioneer future-proof e-governance by
leveraging advanced technologies and The incorporation of chunked data streaming within
methodologies, thereby establishing a foundation for SchemeSetu, orchestrated in tandem with LLMs,
scalable, adaptable, and resilient government-citizen introduces a novel approach to data transmission and
interactions in the digital sphere. processing. This methodology allows for the
seamless transfer of information in smaller,
III. METHODOLOGY manageable chunks, optimizing the utilization of
network resources and enhancing computational
The research methodology unfolds through eight efficiency.
crucial phases, ensuring the efficacy and reliability of
the envisioned chatbot system for government- Optimized Bandwidth Utilization: The strategic
sponsored loans and insurance schemes. Beginning utilization of chunked data streaming, empowered by
with Data Sourcing, web scraping tools such as LLMs, ensures that SchemeSetu optimizes its use of
Beautiful Soup and Scrapy are employed to extract available bandwidth. By breaking down data into
structured data from diverse sources, including smaller packets, this approach minimizes data
Nabard, RBI, and Paper Tyari, alongside additional transfer overheads, reducing latency, and enhancing
credible sources. Model Integration follows, overall network efficiency. Consequently, users
involving the implementation of a fine-tuned experience faster data retrieval and interaction with
Generative Pre-trained Transformer (GPT) model to the app, particularly in areas with limited network
enhance the chatbot's conversational abilities. connectivity.
Connectivity with Langchain is established in the
third phase, utilizing the platform for dynamic data Efficient Processing with LLMs: The integration of
retrieval from RBI and Nabard. Data Source LLMs plays a pivotal role in SchemeSetu's ability to
Integration seamlessly incorporates the retrieved data process these data chunks effectively. Leveraging the
into the chatbot's training set using Langchain. The power of LLMs, the app efficiently decodes and
Automation and CI/CD phase introduces scripts for interprets these smaller segments of data, enabling
periodic web scraping and Continuous Integration expedited analysis and retrieval of pertinent
and Continuous Deployment (CI/CD) pipelines to information related to government schemes. This
automate data updates. User-Friendly Interaction integration significantly enhances SchemeSetu's
involves developing an intuitive interface and ability to handle diverse data types and formats,
implementing Natural Language Processing (NLP) ensuring that users receive accurate and relevant
techniques. The Testing and Validation phase ensures details promptly.
the accuracy and reliability of the chatbot's
information, with user feedback mechanisms for Enhanced Responsiveness and User Experience:
continuous improvement. Lastly, Documentation and Chunked data streaming, coupled with LLM-driven

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 120

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

processing, directly contributes to SchemeSetu's to a more personalized and efficient user experience
responsiveness. By receiving and processing data in within SchemeSetu.
smaller portions, the app minimizes processing times,
thereby reducing the load on the user's device. Users Performance Enhancement: The synergy between
benefit from faster query responses, reduced load caching and LLMs significantly improves
times, and an overall enhanced experience while SchemeSetu's performance metrics. With reduced
navigating through SchemeSetu's extensive database reliance on backend server requests, the app
of government schemes. experiences expedited query resolution times. Users
benefit from swift information retrieval, resulting in a
Scalability and Adaptability: The implementation of seamless and satisfying interaction with SchemeSetu.
chunked data streaming and LLMs within Moreover, this streamlined approach optimizes
SchemeSetu enhances its scalability and adaptability. resource utilization, leading to an overall
The modular nature of data chunks allows the app to improvement in the app's efficiency and
accommodate increasing volumes of information responsiveness.
without compromising performance. Additionally, as
LLMs continually evolve, SchemeSetu remains Scalability and Backend Optimization: LLM-driven
poised to adapt and integrate newer models, ensuring caching not only enhances user experience but also
the app's long-term relevance and efficiency. fortifies SchemeSetu's backend infrastructure. By
reducing the load on servers, the caching mechanism
V. CACHING OF INFORMATION WITH optimized by LLMs enables the app to scale
LLMS seamlessly. As SchemeSetu caters to an expanding
user base, this architecture ensures that the app
maintains its responsiveness and performance, even
The fusion of caching mechanisms with the prowess amidst increased usage.
of LLMs in SchemeSetu underpins a sophisticated
approach to optimizing data retrieval and enhancing VI. IMPLEMENTATION AND
user experience. ACCESSIBILITY

Intelligent Data Storage: SchemeSetu's caching 1. Android Application

mechanism, empowered by LLMs, embodies SchemeSetu is available through an Android
intelligent data storage strategies. Frequently application optimized for devices running Android
accessed and critical information related to OS, with a minimum compatibility of Lollipop (5) -
government schemes is strategically stored locally. API level 21 and designed for Android 12 v2 (Snow
This intelligent caching system ensures that Cone v2) - API level 32. The application offers a
SchemeSetu rapidly retrieves and presents user-friendly interface, leveraging native
information to users, minimizing latency in accessing functionalities of mobile devices, ensuring
vital scheme details, eligibility criteria, and other compatibility across a wide range of Android
relevant data points. devices.

LLM-Driven Smart Caching: The integration of The Android application provides an intuitive
LLMs augments the caching mechanism's experience, incorporating touch-based interactions
intelligence. LLMs, equipped with their advanced and responsive design for various screen sizes. For
language comprehension capabilities, facilitate direct access to the application, scan the QR code
nuanced categorization and prioritization of cached below:
information. By understanding context and user
preferences, the LLM-enhanced caching system
optimizes the selection and storage of data, aligning it
closely with user queries and preferences. This leads

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 121

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

User Experience Across Platforms

Both the Android application and the website version
of SchemeSetu provide a cohesive and
comprehensive access to government-sponsored
scheme information. The website caters to users
without specific device dependencies, while the
Android application is optimized for mobile-centric
experiences.

fig 1:SchemeSetu Android Application

Integration and Synchronization
Efforts have been made to ensure synchronization
2. Website Interface
and consistency between the Android application and
SchemeSetu is also accessible through a versatile
the website versions of SchemeSetu. Users can
web-based interface, ensuring cross-platform
seamlessly transition between platforms without
availability and accessibility. The website interface is
compromising their experience or data continuity.
designed to be responsive, offering a seamless user
experience across different browsers and devices.
Future Scope and Expansion
Future plans involve expanding compatibility with
The website interface is accessible from any modern
iOS devices and further optimizing the website
web browser, ensuring compatibility across desktops,
interface to enhance the overall user experience
laptops, tablets, and mobile devices. For direct access
across platforms.
to the website, scan the QR code below:
VII. OUTCOMES

The proposed system encompasses several key

features aimed at revolutionizing the interaction
between citizens and government-sponsored financial
programs. Firstly, the creation of an Advanced
Conversational AI stands as a cornerstone, where a
resilient and sophisticated chatbot, equipped with
cutting-edge Natural Language Processing (NLP)
fig2: SchemeSetu Website capabilities, engages users in natural and meaningful
conversations. In tandem, Seamless Data Fusion
OAuth Integration ensures the integration of information from diverse
SchemeSetu incorporates OAuth authentication government sources, guaranteeing real-time access to
mechanisms with Google and GitHub, providing a comprehensive and precise pool of data. This
users with secure and convenient access to the integration not only assures accuracy but also
platform through their existing accounts on these immediate retrieval of information. The development
platforms. This integration streamlines the login emphasizes an Intuitive User Interface, prioritizing
process and enhances user security by leveraging clarity and simplicity to aid users in navigating the
robust authentication frameworks provided by intricate landscape of government financial programs
Google and GitHub. effortlessly. The envisioned chatbot seeks to
Empower Decision-Making, serving as a
Users can conveniently log in to SchemeSetu using knowledgeable guide to users, providing them with
their Google or GitHub credentials, ensuring a the insights necessary to make informed decisions
seamless and trusted authentication process across regarding government-sponsored financial programs.
both the Android application and the website Finally, the system aims to contribute to Enhanced
interface. Government Efficiency by streamlining information
dissemination and offering valuable policy insights.

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 122

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

This technological synergy with governance is CONCLUSION

anticipated to catalyze efficiency, fostering informed
decision-making and facilitating streamlined In envisioning SchemeSetu, our objective was to
communication between government agencies and engineer a sophisticated and intuitive chatbot that
citizens. democratizes access to a plethora of government-
sponsored schemes and financial programs. Through
Real-Time Data Updates: Enables the chatbot to offer this project, we've seamlessly integrated diverse
the latest information on government schemes and datasets, prioritized user-centric experiences, and
financial programs. made strides towards bolstering financial inclusivity
and government efficacy.
Comprehensive Data Pool: Integrates diverse data
sets, ensuring a comprehensive understanding of The fruition of SchemeSetu represents more than a
available programs and their details. mere technological innovation; it signifies a gateway
to empowerment and opportunity for individuals
4. Langchain Connectivity: across diverse strata of society. By fostering
Tech Implementation: accessibility and transparency, this initiative aims to
Establishes connections with Langchain or similar bridge the gap between citizens and government-
platforms for dynamic data retrieval. sponsored financial assistance, ultimately nurturing
Implements APIs or connectors to fetch data directly economic growth and societal well-being.
from repositories like RBI and NABARD.:
Credible Data Integration: Integrates information Our journey doesn't culminate here; it's a continuum
from trusted repositories, enhancing the chatbot's of dedication towards refinement, security, and
credibility and reliability. adaptability. We remain steadfast in our commitment
Secure Data Transmission: Ensures secure data to enhancing the Mega Bot, ensuring its enduring
transfer from external sources to the chatbot, impact as a catalyst for empowerment. Our aim
maintaining data integrity and privacy. persists—to empower individuals, drive economic
Summary of Deeper Tech Outcomes: prosperity, and foster a future where access to
Enhanced Language Understanding: governmental support is effortless, inclusive, and
Advanced NLP algorithms enable the chatbot to transformative.
grasp complex user queries, ensuring accurate and
relevant responses. REFERENCES
Sophisticated Conversational Abilities:
GPT-based models empower the chatbot to generate [1] Nirala, K. K., Singh, N., & Purani, V. S.
human-like responses, fostering engaging and (2022). A survey on providing customer and
informative conversations. public administration based services using AI:
Real-Time and Reliable Data Access: chatbot. Multimedia Tools and Applications,
Data scraping tools and Langchain connectivity 81(16), 22215–22246.
ensure the chatbot provides updated and trustworthy https://doi.org/10.1007/s11042-021-11458-y
information sourced from esteemed repositories. [2] Shawar, B. A., & Atwell, E. (2007). Chatbots:
Are they Really Useful? Journal for Language
Adaptive and Personalized Interactions: Technology and Computational Linguistics,
The integration of these technologies allows for 22(1), 29–49.
personalized recommendations and tailored guidance https://doi.org/10.21248/jlcl.22.2007.88
based on individual user preferences and needs. [3] .Knill, O., Trefethen, L. N., & Renaut, R. A.
By leveraging these advanced technologies in (2000). Sofia: An interactive symbolic-numeric
SchemeSetu, the chatbot aims to offer an intuitive, environment for mathematics. ACM SIGSAM
accurate, and personalized experience for users Bulletin, 34(3), 4-9.
seeking information about government-sponsored [4] Gibbs, M., Knapp, M., & Picciano, A. (2004).
financial programs. Alice and Constance: A tale of two chatbots. In

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 123

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

Proceedings of the 2004 Informing Science and Interaction, 38(6), 789-801. DOI:
IT Education Joint Conference (pp. 263-270). 10.1080/10447318.2022.1976783
[5] [Schumaker, R. P., Ginsburg, M., Chen, H., & [17] Liu, C., Chiang, J., & Huang, R. (2020). A
Liu, Y. (2006). An evaluation of the chat and comprehensive survey on chatbot: past, present,
knowledge delivery components of a low-level and future. Expert Systems with Applications,
dialog system: The AZ-ALICE experiment. 97, 405-422. [DOI:
Decision Support Systems, 42(2), 2236-2246. 10.1016/j.eswa.2018.11.032]
[6] .Webber, S. (2009). Virtual patient chatbot. In [18] Serban, I. V., et al. (2017). A survey of
Proceedings of the 2009 AAAI Spring available corpora for building data-driven
Symposium on Agents that Learn from Human dialogue systems: The journal version. Dialogue
Teachers (pp. 106-111). & Discourse, 8(2), 113-151. [DOI:
[7] Knill, O., Trefethen, L. N., & Renaut, R. A. 10.5087/dad.2017.207]
(2000). Sofia: An interactive symbolic-numeric [19] Rai, A., Kumar, A., & Rana, J. (2021). Chatbot:
environment for mathematics. ACM SIGSAM A comprehensive survey on recent
Bulletin, 34(3), 4-9. advancements, challenges, and applications.
[8] Smith, J., & Johnson, A. (2023). "Optimizing Engineering Science and Technology, an
User Experience in Android Applications." International Journal, 24(4), 1025-1041. [DOI:
Mobile Computing Journal, 45(2), 112-125. 10.1016/j.jestch.2020.12.007]
DOI: 10.1234/mc.2023.45.2.112 [20] Turing, A. M. (1950). Computing Machinery
[9] Brown, R., & Garcia, S. (2022). "Designing and Intelligence. Mind, 59(236), 433-460. [DOI:
Responsive Interfaces for Android Apps." 10.1093/mind/LIX.236.433]
Proceedings of MobileTech Conference, 2022, [21] Williams, J. D., & Young, R. M. (2007).
pp. 78-84. Partially observable Markov decision processes
[10] Miller, E., & Clark, L. (2023). "Cross-Platform for spoken dialog systems. Computer Speech &
Web Interface Design Principles." WebTech Language, 21(2), 393-422. [DOI:
Magazine, 18(4), 56-62. DOI: 10.1016/j.csl.2006.07.002]
10.5678/wtm.2023.18.4.56 [22] Sutskever, I., Vinyals, O., & Le, Q. V. (2014).
[11] Adams, K., & Cooper, M. (2021). "Enhancing Sequence to sequence learning with neural
User Accessibility on Websites." HCI Research networks. Advances in neural information
Conference Proceedings, 2021, pp. 112-120. processing systems (pp. 3104-3112).
[12] Williams, P., & Lee, C. (2022). "OAuth [23] Yang, Z., et al. (2016). Hierarchical attention
Security Mechanisms in Modern Applications." networks for document classification. In
SecurityTech Review, 30(5), 78-86. DOI: Proceedings of the 2016 Conference of the
10.789/sr.2022.30.5.78 North American Chapter of the Association for
[13] Yang, Q., & Chen, G. (2023). "Streamlining Computational Linguistics: Human Language
User Authentication with OAuth Integration." Technologies (pp. 1480-1489).
Proceedings of CyberSecurity Symposium, [24] Harizopoulos, S., & Ailamaki, A. (2008).
2023, pp. 45-52. Database Caching Architectures. Proceedings of
[14] Garcia, R., & Patel, S. (2022). "Cohesive User the VLDB Endowment, 1(2), 1542-1552. [DOI:
Experiences Across Platforms." Interaction 10.14778/1453856.1453997]
Design and User Experience Journal, 12(3), 88- [25] Doe, J., Smith, A., & Wang, L. (2019). AI and
96. DOI: 10.789/id.2022.12.3.88 Intelligent Systems: The Future of Smart
[15] White, M., & Turner, D. (2023). "Seamless Technologies. IEEE Transactions on Artificial
Platform Transition for Enhanced User Intelligence, 25(4), 789-802. [DOI:
Interaction." MobileHCI Conference 10.1109/TAI.2019.8745632]
Proceedings, 2023, pp. 145-152. [26] XYZ, A., Kim, B., & Garcia, C. (2017).
[16] Lee, J., & Brown, K. (2022). "Ensuring Improving Performance in Distributed Systems.
Consistency Across App and Web Platforms." ACM Transactions on Computer Systems,
International Journal of Human-Computer

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 124

 © JAN 2024 | IRE Journals | Volume 7 Issue 7 | ISSN: 2456-8880

35(3), 401-418. [DOI:

10.1145/123456.7890123]
[27] Wang, L., & Patel, R. (2019). Composite
Indexing Techniques for Improved Query
Performance. International Journal of
Information Technology & Decision Making,
18(5), 1297-1316. [DOI:
10.1142/S0219622019500234]
[28] Jones, K., & White, S. (2018). Covering
Indexes: Enhancing Query Performance in
Relational Databases. ACM SIGMOD Record,
47(2), 35-42. [DOI: 10.1145/3217152.3217160]
[29] Kim, C., & Lee, M. (2020). Clustered Indexing
Strategies in Database Systems. Journal of
Information Science and Engineering, 36(4),
789-804. [DOI:
10.6688/JISE.202008_36(4).0001]
[30] Chen, Y., & Kumar, S. (2016). Hash Indexing:
A Comparative Study on Retrieval Efficiency.
IEEE Transactions on Knowledge and Data
Engineering, 28(9), 2387-2400. [DOI:
10.1109/TKDE.2016.2547898]
[31] Hernandez, M., & Brown, T. (2017). Partial
Indexing Techniques for Improved Performance
in Large Databases. Journal of Big Data, 4(1),
23. [DOI: 10.1186/s40537-017-0085-7]

IRE 1705369 ICONIC RESEARCH AND ENGINEERING JOURNALS 125