AIClerk: A Multi-Agent Emotion-Aware and Personalized Smart Office Assistant with WhatsApp Integrati by IRJET Journal

AIClerk: A Multi-Agent Emotion-Aware and Personalized Smart Office Assistant with WhatsApp Integration and Automated Academic Workflow Handling

Sayed Mohd Zayeem Khateeb, Javeed Jatagar, Sudeep Kabadagi, Sachin Pattar, Prof. Shradha H

Sayed Mohd Zayeem Khateeb, Student, Dept. of AI & DS, Angadi Institute of Technology and Management, Belagavi, Karnataka, India

Javeed Jatagar, Student, Dept. of AI & DS, Angadi Institute of Technology and Management, Belagavi, Karnataka, India

Sudeep Kabadagi, Student, Dept. of AI & DS, Angadi Institute of Technology and Management, Belagavi, Karnataka, India

Sachin Pattar, Student, Dept. of AI & DS, Angadi Institute of Technology and Management, Belagavi, Karnataka, India

Prof. Shradha H, Assistant Professor, Dept. of AI & DS, Angadi Institute of Technology and Management, Belagavi, Karnataka, India ***

Abstract - The increasing demand for automation and intelligent assistance in academic and professional environments calls for AI systems that are not only functional but also emotionally intelligent, personalized, and real-time accessible. This paper presents AIClerk, a smart office assistant designed using a multi-agent architecture that combines naturallanguageunderstanding,emotiondetection, taskplanning, memory-drivenpersonalization,andWhatsApp integration. The proposedsystemaccepts input via text,voice, or WhatsApp andprocesses it usingadvancedNLPtechniques. It detects user emotions using a transformer-basedmodeland adjusts task responses accordingly. A memory module personalizes task execution based on user history, while an auto-correction system ensures task reliability. Additionally, AIClerk introduces an innovative solution for academic institutions a QR-based exam fee tracking and department notification system that automatically processes student fee submissions via screenshot uploads. Withreal-worldusability and deep AI integration, AIClerkoffers a unique, scalable, and human-like AI assistant for academic and administrative environments.

Key Words: ArtificialIntelligence,SmartOffice Assistant, Emotion Detection, Multi-Agent System, Natural Language Processing, Task Automation, WhatsApp Integration, Personalized AI, Academic Workflow, AIClerk

1.INTRODUCTION

Theincreasingrelianceondigitalsystemsinacademicand organizationalenvironmentshascreatedastrongdemand forintelligent,reliable,anduser-friendlyautomationtools. Traditionalofficeassistantsandtaskmanagersoftenrelyon rule-based logic with minimal adaptability or personalization. Moreover, most of these systems lack emotionalsensitivity,contextualawareness,and real-time

communicationflexibility,whichareessentialforhandling dynamicandhuman-centricworkflows.

Toaddresstheselimitations,wepropose AIClerk  asmart office assistant built on a multi-agent architecture that leverages Natural Language Processing (NLP), emotion recognition,memory-drivenpersonalization,andreal-time messagingintegration.AIClerk notonlyinterpretsnatural languageinstructionsbutalsoadjuststaskbehaviorbased on user mood, past habits, and execution context. By integrating modern AI techniques with practical task automation,thesystemcreatesaresponsiveandhuman-like interactionexperience.

Akeyinnovationinthisprojectistheinclusionofacademic workflow automation, specifically a feature that allows studentstouploadproofofexamfeepayment(withUTRand screenshot)viatheinterface.Thesystemthenautomatically detects the student’s identity and informs the respective department reducing administrative overhead and manualverificationtasksincolleges.

Furthermore, AIClerk incorporates a WhatsApp-based interactionlayerthatmakesitaccessibletousersonmobile devices without needing a separate app. This increases usability and ensures that essential tasks and alerts are deliveredinrealtime,whereusersaremostactive.

This paper outlines the design, implementation, and deploymentofAIClerk,highlightingitsmodulararchitecture, technicalcomponents,andpotentialreal-worldapplications inacademicinstitutionsandbeyond.

1.1 Background and Motivation

In recent years, the integration of artificial intelligence intoofficeandacademicworkflowshasbecomeincreasingly vital for enhancing productivity, reducing repetitive tasks,

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

andimprovinghuman-computerinteraction.Traditionaltask automationtoolsandchatbotsareoftenlimitedtobasicrulebased systems that lack personalization, adaptability, and contextualunderstanding.

Moreover,modernofficeandacademicsettingsdemand toolsthatarenotonlyintelligentbutalsoemotionallyaware and responsive to user behavior. Students and staff frequently interact with digital systems for scheduling, communication, reminders, and administrative tasks. However, these interactions often occur through rigid platforms that do not offer natural, conversational, or emotion-sensitiveexperiences.

ThemotivationfordevelopingAIClerkstemsfromrealworldchallengesobservedinacademicinstitutions,suchas repetitive manual processing of exam fee submissions, delayed communications, lack of timely task updates, and difficulty accessing centralized systems. These problems inspired the need for a human-like, AI-powered assistant capable of understanding user emotions, remembering preferences, communicating over familiar platforms like WhatsApp,andautomatingworkflows.

AIClerkaimstobridgethisgapbyofferinganintelligent, adaptive,andpersonalizedassistantthatstreamlinesdaily office and academic tasks, ultimately improving efficiency andusersatisfaction.

1.2 Limitations of Existing Systems

Despite the increasing use of automation and AI-based assistantsinofficeandacademicsettings,existingsystems have several limitations that affect their practical effectivenessandusersatisfaction:

 Lack of Emotion Awareness: Mostvirtualassistantsandautomationtoolsarenot capable of detecting or adapting to the user's emotional state. As a result, they deliver generic responsesthatmaynotalignwiththeuser'smood orurgency.

 Absence of Personalization and Memory: Conventionalsystemsdonotretainanymemoryof previous interactions, user preferences, or frequentlyperformedtasks.Thisleadstorepetitive inputsandalackofpersonalizeduserexperience.

 Limited Adaptability in Task Execution: Current solutions operate based on predefined workflowsorrule-basedlogic,whichcannotadjust dynamicallytochangesinusercontext,sentiment, ortaskpriority.

 Inadequate Real-Time Interaction: Many systems are restricted to web-based or desktopinterfacesanddonotsupportmobile-first platformslikeWhatsApp,whichlimitsaccessibility andreducesuserengagement.

 Lack of Support for Academic Workflows: Most existing models are not designed to handle academic-specific tasks such as exam fee verification, department-level notifications, or backlog identification, resulting in continued relianceonmanualadministrativeprocesses.

These limitations emphasize the need for a smarter, emotionallyadaptive,personalized,andmobile-accessibleAI solutionlikeAIClerk.

1.3 Objectives of the Study

The primary objective of this study is to design and implement a smart, AI-powered office assistant named AIClerk thatenhancesuserexperiencebyintegrating emotional intelligence, personalization, and real-time accessibility. The system is tailored for academic and administrative environments to improve productivity, reducemanualoverhead,andprovideaninteractive,usercentricinterface.

Thespecificobjectivesofthestudyareasfollows:

 To develop a multi-agent AI system capable of understanding natural language commands and managingtasksintelligently.

 To implement an emotion detection module that analyzesuserinputandadjuststaskbehaviourand responsetoneaccordingly.

 Todesignamemory-drivenpersonalizationengine thatremembersuserpreferences,pastbehaviours, and frequently performed tasks to improve accuracyandefficiency.

 To build an auto-correction and alert mechanism that detects execution errors, retries tasks, and providesuserfeedbackoralternativesuggestions.

 To integrate the assistant with WhatsApp using APIstoenablemobile-firstinteractionandreal-time tasknotifications.

 Tointroduceauniqueacademicworkflowfeature wherestudentscanuploadexamfeepaymentproof, enabling the system to identify the student and automaticallynotifytherelevantdepartment.

 To provide a unified and accessible platform that simplifies routine academic and office operations throughintelligentautomation.

These objectives aim to demonstrate how artificial intelligence can be effectively applied to solve practical problemsinacollegeorinstitutionalenvironment,offeringa blendofefficiency,adaptability,anduser-friendlydesign.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

1.4 Scope and Contribution

Thisprojectfocusesonthedevelopmentofasmart,multiagentAIassistant  AIClerk thatiscapableofperforming personalized and emotion-aware task automation in academicandofficeenvironments.Thesystemisdesignedto functionasavirtualassistantthatnotonlyunderstandsand executes tasks but also adapts to user behavior and emotional tone to improve user experience and overall efficiency.

Thescopeofthisprojectincludes:

 Designingamodular,AI-drivensystemthataccepts user inputs through text, voice, or WhatsApp and performstaskexecutionthroughintelligentagents.

 Implementing natural language understanding (NLU), emotion detection, and memory-based personalization using AI models and vector databases.

 Automatingofficeandacademicworkflowssuchas meeting scheduling, report generation, and task reminders.

 Creating a dedicated feature for academic institutions to manage exam fee submissions via QR-based screenshot uploads, with automatic identificationanddepartmentnotifications.

 Providing multi-platform accessibility, including WhatsAppintegration,toensurereal-time,mobilefriendlyinteractions.

Thekeycontributionsofthisworkinclude:

 A novel integration of emotion detection and memory-based personalization in a smart office assistant.

 Theuseofamulti-agentarchitecturetobreakdown tasksandexecutethemefficientlyusingspecialized sub-agents.

 Real-world applicability through WhatsApp interaction,whichmakesthesystemaccessibleto students,faculty,andstaffonmobiledevices.

 Anoriginalacademicworkflowautomationfeature that addresses real-time fee tracking and reduces manualprocessingineducationalinstitutions.

Through these contributions, AIClerk demonstrates the practicalapplicationofartificialintelligenceinsolvingdaily operational challenges in academic and administrative domains.

1.3 Literature Survey

hissectionpresentsasurveyofkeyresearchpapersand projects related to emotion-aware AI systems, personalizationinvirtualassistants,workflowautomation, and messaging platform integrations. While these studies contributedsignificantlytotheadvancementofintelligent assistants,theyalsopresentlimitationsthatthisproject AIClerk aimstoovercome.

Several researchers have explored the development of emotion-aware frameworks and AI agents for domainspecificapplications.However,manyofthesesystemsfocus onindividualaspectssuchaschatbot-basedtherapy,basic intent recognition, or limited automation workflows. A summary of selected papers, the technologies they used, theirdrawbacks,andtheimprovementsintroducedbyour proposedmodelispresentedinTable1.

Table-1:ComparisonofExistingWorkswithProposed AIClerkModel

Author&Year Technolog y Used Limitations in Existing Work Our Contribution / Improvements

Emotion AWARE, 2022 [1]

AI Personalization ,2021[2]

BERT + LSTM for Emotion Detection

Collaborativ e Filtering, Static Preferences

AutoFlow, 2023[3] Workflow generation usingLLMs

Therapy

Chatbot, 2023 [4]

AutoGen, 2023 [5]

WhatsApp Bot with LLM (LLM4Ther apy)

Detects emotionbut doesn’t adapt tasks or interaction style accordingly

Nomemory or learning overtime

No fallback or autocorrectionif taskfails

Limited to mental health; no practical task automation

Multi-agent LLM framework Generic agents, no emotion or memory integration

Emotion-aware taskhandlingand tone adjustment using fine-tuned transformer model

Memory-based personalization using vector embeddings and userhistory

Auto-correction and alert system witherrortagging andretrylogic

WhatsApp-based interaction for real-world office andacademictask execution

Customized agents for task planning, emotion,memory, and platform communication

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

2. METHODOLOGY

The proposed system, AIClerk, is designed using a modular multi-agent architecture powered by Artificial Intelligence (AI), Natural Language Processing (NLP), and integration APIs. The methodology involves a sequence of intelligentagentsworkingtogethertohandleusercommands inahuman-like,adaptive,andreliablemanner.Eachmodule uses specific technologies and coding implementations to performitsrole.

2.1 Input Collection and Preprocessing

UsersinteractwithAIClerkthroughthreechannels:Web Interface,VoiceCommands,andWhatsAppChat.Thesystem firstcapturestheinputandappliespreprocessingstepssuch as:

 Tokenization

 Lemmatization

 Removalofstopwords

Iftheinputisinvoiceformat,itisconvertedtotextusing the Whisper Speech-to-Text API or Google Speech RecognitionAPI.

-Technologies Used: Python (NLTK, SpaCy), Whisper / GoogleSpeechAPI,Flaskforinterfaceinputrouting

2.2

Emotion Detection Module

Thismoduledeterminestheuser’semotionalstate(e.g., stressed,neutral,happy)basedontheirmessage.Thisisdone usingapre-trainedtransformermodelsuchas:

 BERT

 DistilBERT

 LSTMforvoiceemotionfeatures(optional)

Theresulthelpsthesystemadjustthetask’surgencyor tone.

-Technologies Used: HuggingFace Transformers (BERT/DistilBERT), PyTorch, Scikit-learn, Custom classificationlogic

2.3

Figure 1- Emotiondetection

Task Understanding (Intent Recognition)

This module identifies the purpose behind the user's input(e.g.,“SendreporttoHR,”“Remindmetomorrow”).It extracts:

 Intent

 NamedEntities(e.g.,names,dates,times)

 Context

Weuseafine-tunedLLM(e.g.,OpenAIGPT-3.5orCohere) andNamedEntityRecognition(NER)models.

-Technologies Used: OpenAI API / Cohere API, HuggingFaceTransformers,spaCyNER

2.4 Planner Agent

This agent determines how to complete the task by generatingalogicalsequenceofsub-tasksandassigningthem toappropriateagents(execution,alert,ormemory).

 Generates multi-step workflows based on input complexity

 Usesif-elselogicandcontextawareness

-Technologies Used: Python (custom rule engine), AutoFlow-styleplanning,JSON-basedtasktemplates

2.5 Memory-Driven Personalization Agent

This agent retrieves past user interactions to adapt responses.Forexample,ifauseralwaysprefersreportsas PDFs,itremembersandauto-appliesthis.

 Storesuserhabits/preferencesusingembeddings

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net

 Stores metadata such as frequency of tasks and preferredtimes

-Technologies Used: FAISS or Chroma (vector DB), SQLite/JSON for local history, SentenceTransformers for embeddings

2.6 Task Execution Agent

Thisisthecoremodulethatperformsthedesiredaction, suchas:

 Sendinganemail

 Schedulingatask

 Generatingareport

 Organizingafile

-TechnologiesUsed:Python(smtplibforemails,openpyxl for Excel), Google Calendar API, Pandas for data handling, Flaskforrouting

2.7 Auto-Correction and Alert System

This agent checks whether the task was executed successfully.Ifnot,iteither:

 Retrieswithfallbacklogic

 Suggestsalternativepaths

 Sendsreal-timealerts

The system also logs the error with emotion tag (e.g., errorduringstress).

-Technologies Used: Python (try-except handlers), customloggingfunctions,alertsystemviaTwilio/MailAPI.

2.8 WhatsApp Integration

Users can interact with AIClerk directly through WhatsApp.Thesystemreceivesthemessage,processesthe task,andsendstheresponseback.

 Incomingmessagesarehandledviawebhook

 Task progress and alerts are sent as real-time messages

-Technologies Used: Twilio or 360Dialog API, Flask/Djangobackend,WhatsAppBusinessPlatform

2.9 Multi-Agent Orchestration

Allaboveagentsarecoordinatedbyacentralcontroller.It decides:

 Whichagentshouldrun

 Inwhatorder

 Howtopassdatabetweenthem

Thisensuresthesystembehaveslikeateamofintelligent assistantsworkingtogether.

-TechnologiesUsed:Python(modularclassesforagents), AutoGen-inspiredarchitecture

Figure 2-ConversationbetweenuserandAIClerk

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

3- Multi-Agentsystemflow

2.10 Response Generation and Logging

Oncethetaskiscompleted,thesystem:

 Formsanaturallanguageresponse

 Logs the task, emotion, time, and result for future use

This ensures traceability and helps in memory-based learning.

-Technologies Used: OpenAI / GPT for response generation, SQLite/JSON for log storage, Markdown/Text formattingforreplies

Figure 4- Responsegenerationandlogging

2.11 WhatsApp Payment Automation for Exam Fee Verification (Optional Feature)

Toaddressthecommonproblemofexamfeesubmission delays and verification chaos in academic institutions, AIClerkintroducesauniqueacademicworkflowmodulethat

automates student fee verification and department notificationsusingWhatsAppandOCR.

Theprocessbeginswhentheuniversityannouncesexam feepaymentdeadlines.Studentspaythefeesviathecollege's provided QR code and are then prompted to upload a screenshot of their payment receipt (containing UTR number)throughtheAIClerkinterfaceorWhatsApp.

Oncethescreenshotisreceived,thesystemperformsthe followingsteps:

 Extracts student details (Name, USN, UTR) using OpticalCharacterRecognition(OCR).

 Verifies the format and ensures UTR number is present.

 IdentifiesthedepartmentbasedontheUSN.

 Automatically sends a confirmation alert to the respectivedepartment,indicatingthatthestudent hassuccessfullypaidthefeeandiseligibleforform filling.

Thissystemhelpsreducemanualpaperwork,errorsinfee verification, and backlogs in form processing. It provides transparency to both students and administrators while streamlininginstitutionaloperations.

Technologiesused:

 EasyOCRorTesseractforimage-to-textconversion

 WhatsAppAPIviaTwilio/360Dialogforscreenshot collection

 Python(Pillow,OpenCV)forimagepreprocessing

 SQLite/JSONfordataloggingandtracking

Figure 5- Paymentconfirmationnotificationtodepartment

3. User Interface Designing

Theuserinterface(UI)ofAIClerkisdesignedtobesimple, intuitive, and accessible across both desktop and mobile platforms.Thesystemsupportsinputthrough webforms,

Figure

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

WhatsApp chat, and voice, making it flexible for academic and office use. The interface ensures seamless task interaction,notificationupdates,andscreenshotuploads.

3.1 Frontend

The frontend is responsible for capturing user input, displaying responses, and providing a user-friendly dashboardforstaffandstudents.Itincludeswebpagesand visualcomponentsthatguideuserinteraction.

TechnologiesandLanguagesUsed:

 HTML5–Forbasicpagestructure

 CSS3–Forstylingandlayout

 JavaScript–Forinteractivityandvalidation

 Bootstrap – For responsive design and mobile compatibility

 React.js (optional) – For creating dynamic componentsandscalablefrontendstructure

OptionalChannels:

 WhatsApp Chat Interface (via Twilio API) – for mobile-first interaction with no custom app required

3.2 Backend

Thebackendhandleslogicprocessing,APIintegration,data flow, emotion detection, memory management, and task execution. It connects the UI with all AI services and managestheinternalcommunicationbetweenagents.

TechnologiesandLanguagesUsed:

 Python–Corebackendlogicandmodelintegration

 Flask/Django – Webframework tohandleHTTP requestsandroutes

 SQLite/JSON–Forlightweightlocalstorage

 FAISS / Chroma – For memory-driven personalizationusingvectorembeddings

 HuggingFaceTransformers–ForBERT/DistilBERTbasedemotionrecognitionandNLP

 Twilio / 360Dialog API – For WhatsApp message handling

 Tesseract / EasyOCR – For extracting data from uploadedscreenshots

3. CONCLUSIONS

TheAIClerksystemdemonstratesapracticalandinnovative applicationofartificialintelligenceinstreamliningacademic and office workflows. By integrating emotion detection, intent recognition, memory-based personalization, task automation, and WhatsApp communication into a unified multi-agentframework,thesystemaddressesthelimitations oftraditionaldigitalassistants.

ThroughtheuseofadvancedNLPmodels,vectormemory databases,andreal-timemessagingAPIs,AIClerknotonly enhancesuser experiencebutalsoreducesadministrative burden,improvestaskreliability,andsupportspersonalized taskexecution.Theadditionofasmartacademicmodule capableofhandlingexamfeeverificationanddepartmental notifications further shows the system’s relevance and usabilityininstitutionalenvironments.

Overall, AIClerk offers a scalable, mobile-accessible, and intelligentassistantsolutionthatcanbeadaptedtovarious organizationalneeds,bridgingthegapbetweenuser-friendly interactionandcomplextaskautomation.

ACKNOWLEDGEMENT

We would like to express our sincere gratitude to our respected guide, Prof. Shraddha H, Assistant Professor, Department of Artificial Intelligence and Data Science, AngadiInstituteofTechnologyandManagement,Belagavi, for her continuous guidance, motivation, and valuable insightsthroughoutthedevelopmentofthisproject.

WealsoextendourthankstotheHeadoftheDepartment, projectco-ordinatorandallfacultymembersforproviding the necessary resources and a supportive academic environment.

Figure 6- Frontendandbackend

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

Finally, we thank our friends and family for their encouragementandfeedback,whichhelpedusimprovethe qualityanddirectionofourwork.

REFERENCES

[1] A.Panda,A.Bera,andD.Manocha,“EmotionAWARE:An AIframeworkforadaptableemotionrecognition,” IEEE Trans. Affective Comput., vol. 13, no. 2, pp. 469–482, 2022.

[2] [2]X.LiandJ.Xu,“AI-drivenpersonalizationindigital marketing,” ACMComput.Surv., vol.54,no.3,pp.1–28, 2021.

[3] [3] H. Zhang, K. Wu, and A. Dubey, “AutoFlow: AutomatedworkflowgenerationforLLMagents,” arXiv preprintarXiv:2308.00032,2023.

[4] [4]M.Cohen,T.Levy,andD.Israel,“Narrativetherapy chatbot on WhatsApp using LLM,” Comput. Human Behav.Rep., vol.9,p.100216,2023.

[5] [5]R.Songetal.,“AutoGen:Enablingmulti-agentLLM collaboration,” AutoGen Framework, Microsoft Research, 2023. [Online]. Available: https://github.com/microsoft/autogen

[6] [6]J.Devlin,M.Chang,K.Lee,andK.Toutanova,“BERT: Pre-training of deep bidirectional transformers for languageunderstanding,”in Proc.NAACL-HLT, 2019,pp. 4171–4186.

[7] [7]S.HochreiterandJ.Schmidhuber,“Longshort-term memory,” NeuralComput., vol.9,no.8,pp.1735–1780, 1997.

[8] [8] A. Radford et al., “Language models are few-shot learners,” OpenAI GPT-3 Paper, arXiv preprint arXiv:2005.14165,2020.

[9] [9] K. Xu et al., “Emotion recognition from text using LSTM and attention,” IEEE Access, vol. 9, pp. 11064–11075,2021.

[10] [10]A.Vaswanietal.,“Attentionisallyouneed,”in Adv. NeuralInf.Process.Syst., 2017,pp.5998–6008.

[11] [11]R.Zhang,Y.Sun,andB.Liu,“Personalizedvirtual assistant design with memory graphs,” Springer Artif. Intell.Rev., vol.55,pp.143–161,2022.

[12] [12] Twilio Inc., “Twilio WhatsApp Business API documentation,” TwilioDeveloperDocs, 2024.[Online]. Available:https://www.twilio.com/docs/whatsapp

[13] [13]M.Brown,A.Singh,andL.Patel,“OCR technology for automated data capture in smart apps,” J. Intell. Syst., vol.29,no.4,pp.412–426,2020.

[14] [14] Meta AI Research, “Chroma: Open-source vector database for LLM memory,” 2023. [Online]. Available: https://www.trychroma.com

[15] [15]PythonSoftwareFoundation,“WorkingwithFlask, SQLite, and JSON for AI workflows,” Python.org, 2023. [Online].Available:https://docs.python.org

BIOGRAPHIES

Sayed Mohd Zayeem Khateeb iscurrentlypursuingaB.E.inArtificialIntelligenceand Data Science at Angadi Institute of Technology and Management,Belagavi.Heispassionateaboutbuilding intelligent AI systems, with interests in natural languageprocessing,automation,andproductdesign. Heledtheteam,focusedon methodology,WhatsApp integration,OCR-basedscreenshotuploadmoduleand overall system architecture development for this project.

Javeed Jatagar isanundergraduatestudent intheDepartment ofAI and DS at Angadi Institute of Technology and Management. His areas of interest include machine learning,webdevelopment,andreal-worldapplication design.Hecontributedtothefront-endinterface.

Sudeep Kabadagi is currently pursuing B.E. in AI and Data Science at AngadiInstituteofTechnologyandManagement.His interests lie in data structures, backend automation, andUIdevelopment.Heassistedwithbackendlogic

Sachin Pattar is a final-year student specializing in Artificial Intelligence and Data Science. His primary areas of interestaredeeplearningandappliedAI.Heworked onmodelintegrationandoveralldocumentation

Prof. Shradha H isanAssistantProfessorintheDepartmentofArtificial Intelligence and Data Science at Angadi Institute of TechnologyandManagement,Belagavi.Shehasbeen actively mentoring students in AI, data science, and emerging technologies. She guided the team throughout the project, providing critical insights in researchmethodologyandimplementation.