WhatsApp-Native Eight-State Conversational Bot for Contract Lifecycle Automation in SME Environments

WhatsApp-Native Eight-State Conversational Bot for Contract Lifecycle Automation in SME Environments

Aakash Girhe (1), Dr. Prakash Kene (2)

(1) MCA Student, PES Modern College of Engineering, Pune, India

(2) Assistant Professor, MCA Department, PES Modern College of Engineering, Pune, India

Abstract – Contract lifecycle management (CLM) in small and medium-sized enterprises (SMEs) remains a fragmented, manual process relying on disconnected tools for proposal creation, client communication, document generation, and signature collection. This paper presents Smart CLM, a self-hosted full-stack contract lifecycle management system featuring a novel WhatsApp-native eight-state conversational bot architecture for end-to-end pro-posal initiation and client approval workow. The system further integrates a coordinate-based dynamic PDF template lling mechanism, a parallel electronic signing architecture enabling simultaneous independent signature collection from multiple parties, and cryptographic audit certicate generation compliant with Indias Information Technology Act 2000 and Amendment Act 2008. Implemented using Spring Boot 3.5, React 18, Post-greSQL 15, Python ReportLab, and Evolution API, the system demonstrates a 95.2% reduction in contract execution time compared to traditional manual workows. Evaluation across 67 black-box test cases conrms 100% functional correctness. The platform provides a legally compliant, open-source alternative to enterprise CLM solutions inaccessible to SMEs due to cost and complexity barriers.

Index TermsContract Lifecycle Management, WhatsApp

Business API, Electronic Signature, Digital Certicate, PDF Tem-plate Automation, IT Act 2000, Parallel Signing, State Machine, SME Workow Automation, Spring Boot

1. Introduction

   Contract management is a foundational operational process in every commercial engagement. Despite its importance, the majority of small and medium-sized enterprises (SMEs) in India manage contract workows through a combination of word processors, email, standalone e-signing tools, and shared le storage each operating independently without integration [1].

   The fragmentation creates measurable inefciencies. The International Association for Contract and Commercial Man-agement (IACCM) estimates that organizations lose approxi-mately 9.2% of annual revenue due to poor contract manage-ment [1]. For SMEs and independent freelance professionals operating without dedicated legal teams, this overhead is proportionally more severe.

   A challenge specic to the Indian business context is the mismatch between existing CLM platform communication channels and actual business preferences. WhatsApp has over 500 million active users in India and serves as the dominant channel for professional communication among SMEs [2]. However, no existing CLM platform including Ironclad,

   DocuSign CLM, Adobe Sign, or PandaDoc provides native WhatsApp integration for proposal dispatch, client approval, or contract initiation.

   The cost barrier further limits adoption. Enterprise CLM platforms are priced between USD 500 and USD 2,000 per month [3], placing structured contract management entirely beyond the reach of SMEs and individual freelancers.

   This paper presents Smart CLM, which addresses these limitations through four primary contributions:

   • A WhatsApp-native eight-state nite state machine for complete proposal and approval workow

   • A coordinate-based dynamic PDF template lling system supporting unlimited contract types

   • A parallel electronic signing architecture for simultaneous multi-party signature collection

   • Cryptographic certicate generation compliant with the Information Technology Act 2000 and Amendment Act 2008

2. Related Work

   1. Contract Lifecycle Management Systems

      Lindell and Kirsch [4] identied seven phases in the contract lifecycle: request, authoring, negotiation, approval, execution, obligation management, and renewal. Their framework estab-lished the conceptual foundation for modern CLM platform design and continues to be referenced in commercial CLM architecture documentation.

      Bhatt and Mudambi [5] surveyed CLM adoption patterns across 240 organizations and found that adoption among SMEs remained below 12%, primarily due to cost and the mis-match between platform capabilities and actual SME workow requirements. Kumar and Singh [6] specically studied the Indian IT freelance sector and found that 78% of respondents managed contracts through email and WhatsApp exchanges, with fewer than 8% using any formal CLM tool. The primary barriers cited were cost (67%), complexity (48%), and the absence of WhatsApp integration (71%).

   2. Electronic Signature and Digital Certicates

      Adams and Lloyd [7] established the theoretical foundation for certicate-based signing architectures including public key infrastructure design and RFC 3161 timestamp author-ity protocols. Szyjewski [8] identied three primary attack

      vectors in digital document signing and proposed HMAC-based verication combined with chained hash structures as countermeasures, directly informing Smart CLMs certicate design.

      Chen et al. [9] demonstrated that parallel signing reduces av-erage contract execution time by 61% compared to sequential workows while maintaining equivalent legal validity, provid-ing empirical support for the parallel signing architecture in Smart CLM.

      Verma [10] analyzed the IT Act 2000 requirements for elec-tronic signature validity in India, identifying four minimum technical requirements: a veriable document hash, a reliable timestamp, signatory identity, and a signing certicate. Smart CLMs certicate generation mechanism is designed to satisfy all four requirements.

   3. WhatsApp-Based Business Automation

      Kovtun and Tulyakov [11] analyzed chatbot-based systems for business process management and identied six component groups that determine chatbot effectiveness, particularly em-phasizing robust state management for multi-turn conversation ows.

      Patel and Mehta [12] implemented a WhatsApp-based order management system for Indian SMEs, demonstrating a 45% reduction in processing time and a 67% increase in customer response rates compared to email-only workows. This es-tablishes WhatsApp as a viable primary business transaction channel for the Indian SME context.

      Rodrigues et al. [13] documented the LID (Linked Device Identier) mapping problem that arises in WhatsApp Business API integration, where messages from certain clients carry LID-format identiers differing from the stored phone number format. This problem directly manifests in Smart CLMs bot implementation and motivated the LID mapping persistence mechanism described in Section IV.

   4. PDF Generation and Template Automation

      Bhardwaj and Rathi [14] compared four approaches to dynamic PDF generation and found that overlay-based eld lling provided the highest template delity for existing document layouts, motivating the coordinate-based overlay approach in Smart CLM.

      Marciniak et al. [15] developed a visual eld mapping in-terface for a legal document management system and reported 94% user satisfaction with the visual approach compared to coordinate specication through conguration les.

   5. Research Gap

   The literature reveals four unddressed gaps: (1) no CLM system uses WhatsApp as the primary contract initiation channel; (2) no open-source system generates IT Act 2000 compliant certicates; (3) dynamic coordinate-based template lling is not integrated into end-to-end CLM workows; and

   (4) no unied open-source CLM exists that is cost-accessible to SMEs and freelancers. Smart CLM addresses all four gaps simultaneously.

3. System Architecture

   1. Overall Architecture

      Smart CLM follows a layered architecture comprising six functional layers: the React 18 browser-based frontend, the Spring Boot 3.5 REST API backend, the PostgreSQL 15 relational database, the Python 3 PDF processing module, the Evolution API WhatsApp gateway, and the Gmail SMTP email delivery service. All services are containerized using Docker Compose for consistent deployment.

      Fig. 1. Smart CLM System Architecture

   2. Database Design

      The system uses PostgreSQL 15 with eight primary entities: USER, PROPOSAL, SIGNINGSESSION, CONTRACTTEMPLATE, STOREDDOCUMENT, TASK, WHATSAPPSESSION, and LIDMAPPING. Schema

      management is handled through Flyway migration scripts. The SIGNINGSESSION entity stores both parties tokens, signature images as binary data, the original and nal signed contract PDFs, and the dynamic eld mapping as a JSON string.

   3. Security Architecture

   Authentication is implemented using JSON Web Tokens (JWT) with role-based access control enforcing three permis-sion levels: ADMIN, MANAGER, and STAFF. Signing portal access is controlled through unique 128-bit SecureRandom tokens per session, expiring after a congurable duration.

4. Implementation

   1. WhatsApp Bot Eight-State Machine

      The WhatsApp bot is implemented as an eight-state nite state machine (FSM) persisted in the WHATSAPPSESSION table. Each incoming WhatsApp message triggers the handleMessage() method in WhatsappBotService,

      TABLE I

      WhatsApp Bot State Transitions

      State Trigger Action

      overlay with the source using pypdfs page merge operation. This preserves all original document formatting while injecting dynamic content at specied positions.

      1. Parallel Electronic Signing

         IDLE Greeting

         message

         MENU Option

         number

         PROPOSALTYPE Type

         selected

         Send menu, transition to MENU

         Route to workow state

         Store type, COL-LECTING

         Upon contract creation, the system generates two indepen-dent 128-bit SecureRandom tokens one for the client and one for the developer stored in the SIGNINGSESSION record alongside boolean ags client_signed and dev_signed, initially false.

         COLLECTINGDATA Field input Collect name, email,

         phone

         REVIEW Approve/cancel Send to client or reset CLIENTAPPROVAL Client reply Update status, notify

         staff

         When either party submits their signature through the browser-based signing portal, the signature image is stored as a PNG byte array and the corresponding ag is set to true. The checkAndComplete() method evaluates whether both

         CONFIRMCONVERT Yes/No

         from staff

         SIGNATORYSELECT Developer

         select

         Create signing session

         Send signing links to both

         ags are true after each submission. When both signatures are present, the method triggers nal PDF generation, overlaying both signatures at their congured eld positions and append-ing the cryptographic certicate page.

      2. Cryptographic Certicate Generation

      which retrieves the current state for the senders phone number and delegates to the appropriate state handler.

      A critical implementation challenge was the LID mapping problem. WhatsApp routes messages from certain clients using a LID (Linked Device Identier) format rather than the stan-dard E.164 phone number format. Smart CLM resolves this through a LIDMAPPING table persisting the bidirectional mapping between LID and JID formats, ensuring consistent session retrieval regardless of the identier format in incoming messages.

      Fig. 2. WhatsApp Bot Conversation Proposal Creation Flow

   2. Dynamic PDF Template Filling

      The PDF generation pipeline implements a four-stage process: template retrieval, eld map parsing, overlay gen-eration, and PDF merging. The eld map stored in CONTRACT_TEMPLATE.field_map is a JSON array of eld descriptor objects, each specifying eld identier, page number, x and y coordinates in PDF user-space units, width, height, and eld type (TEXT or SIGNATURE).

      The Python PDF engine creates an invisible overlay PDF at the exact dimensions of each template page, draws text or signature images at the specied coordinates, and merges the

      The certicate generation process implements four opera-tions:

      1. SHA-256 hash of the complete signed PDF byte array computed using Javas MessageDigest

      2. HMAC-SHA512 verication code generated using a

         server-side secret key and concatenation of contract ID, signatory names, and the SHA-256 hash

      3. UTC timestamp recorded as certicate generation time with signatory-level granularity

      4. Certicate page formatted with all data and IT Act 2000 compliance statement, appended to the nal PDF using Python ReportLab

5. Experimental Validation

   1. Functional Testing

      The system was evaluated through 67 black-box test cases across 10 functional modules. All 67 test cases passed, con-rming 100% functional correctness across all system modules as shown in Table II.

      TABLE II

      Functional Test Results Summary

      Module	Total	Pass	Rate
      Authentication & RBAC	8	8	100%
      Proposal Management	8	8	100%
      WhatsApp Bot Workow	8	8	100%
      Contract Management	5	5	100%
      Template Management	6	6	100%
      Electronic Signing	14	14	100%
      Document Storage	4	4	100%
      Task Management	4	4	100%
      Analytics & Reporting	6	6	100%
      Role-Based Access	4	4	100%
      Total	67	67	100%

      Fig. 3. Browser-Based Parallel Signing Portal

      Fig. 4. Generated Cryptographic Certicate Page

   2. Performance Evaluation

      Contract execution time was measured across three work-ow congurations over 20 contract cycles with two signato-ries each, as shown in Table III.

      TABLE III

      Contract Execution Time Comparison

      Stage	Manual	Semi-Auto	Smart CLM
      Proposal creation	45 min	8 min	3 min
      Client approval	24 hrs	4 hrs	30 min
      Contract generation	30 min	5 min	1 min
      Signature collection	48 hrs	12 hrs	2 hrs
      PF delivery	20 min	3 min	1 min
      Total	73.5 hrs	16.5 hrs	3.5 hrs

      Smart CLM reduces total contract cycle time from 73.5 hours (manual) to 3.5 hours a 95.2% reduction. The parallel signing architecture contributes the largest single

      improvement, reducing signature collection from 48 hours (sequential) to 2 hours (parallel), consistent with Chen et al.

      [9] who reported 61% reduction for parallel signing.

   3. Platform Comparison

   Table IV compares Smart CLM against existing commercial CLM platforms across four critical dimensions for Indian SME users.

   TABLE IV

   Comparison With Existing CLM Platforms

   Platform	Price/mo	WA	Parallel	IT Act
   Ironclad	$2,000+	No	Yes	No
   DocuSign CLM	$1,500+	No	Yes	No
   Adobe Sign	$500+	No	Yes	Partial
   PandaDoc	$49+	No	No	No
   Smart CLM	Free	Yes	Yes	Yes

6. Discussion

   Smart CLM is useful because it bridges practical gaps in SME contract workows. The WhatsApp-native bot architec-ture is particularly signicant because it changes the interac-tion model from access a web portal to initiate a contract to send a WhatsApp message. In the Indian SME context, where WhatsApp is the primary business communication tool for over 71% of users [6], this shift dramatically reduces friction in the contract initiation process.

   The parallel signing architecture addresses a real pain point where sequential signing workows create articial delays when both parties are simultaneously available. The coordinate-based template lling system enables organizations to use any existing PDF contract template without modifying source les or rewriting application code, making adoption signicantly simpler than alternative approaches.

   From a legal perspective, the explicit IT Act 2000 com-pliance in the generated certicates addresses a gap not covered by any existing open-source CLM solution, making Smart CLM directly applicable in Indian legal and commercial contexts.

7. Limitations and Future Work

   Current limitations include: (1) PDF eld coordinate cal-ibration requires manual adjustment due to browser-to-PDF scaling differences; (2) the system relies on Evolution APIs unofcial WhatsApp Web connection, which may be disrupted by platform changes; (3) single-language (English) support limits accessibility for regional language users; (4) obligation tracking post-signing is not yet implemented.

   Future work will address these through: automatic scale fac-tor detection for the eld editor; migration to the ofcial Meta WhatsApp Business Cloud API for enterprise-grade reliability; multilingual bot responses in regional Indian languages; AI-assisted contract risk scoring using large language models; and blockchain-based immutable audit trails using Ethereum smart contracts for enhanced legal enforceability.

8. Conclusion

This paper presented Smart CLM, a full-stack self-hosted contract lifecycle management system that addresses fragmen-tation, cost barriers, and communication channel mismatches preventing SMEs and freelance professionals from adopting structured contract management. The system introduces four novel contributions: a WhatsApp-native eight-state bot ar-chitecture, coordinate-based dynamic PDF template lling, parallel electronic signing, and IT Act 2000 compliant cryp-tographic certication.

Experimental evaluation demonstrated a 95.2% reduction in contract cycle time compared to traditional manual workows, and 100% functional correctness across 67 test cases. Smart CLM provides a legally compliant, open-source alternative to enterprise CLM platforms priced beyond the reach of SMEs and individual professionals operating in the Indian market.

References

1. International Association for Contract and Commercial Management (IACCM), The Cost of Poor Contract Management, IACCM Research Report, 2019.

2. Statista Research Department, Number of WhatsApp Users in India from 2015 to 2025, Statista Digital Market Outlook, 2023.

3. G. Saxena and R. Gupta, Comparative Analysis of Enterprise Contract Lifecycle Management Platforms, International Journal of Business Information Systems, vol. 38, no. 2, pp. 145162, 2022.

4. Y. Lindell and B. Kirsch, Contract Lifecycle Management: A Frame-work for Modern Enterprises, Journal of Contract Management, vol. 12, no. 1, pp. 2341, 2015.

5. D. Bhatt and R. Mudambi, Cloud-Based CLM Adoption Patterns Across Enterprise Segments, Journal of Information Technology, vol. 35, no. 3, pp. 201218, 2020.

6. R. Kumar and A. Singh, Contract Management Challenges in the Indian IT Freelance Sector, International Journal of Engineering Research & Technology (IJERT), vol. 11, no. 4, pp. 234241, 2022.

7. C. Adams and S. Lloyd, Understanding PKI: Concepts, Standards, and Deployment Considerations, 2nd ed. Addison-Wesley, 2002.

8. G. Szyjewski, Digital Document Signing: Vulnerabilities and Solu-tions, IEEE Access, 2023.

9. X. Chen, Y. Wang, and Z. Liu, Parallel vs. Sequential Multi-Party Contract Signing: Performance and Legal Analysis, IEEE Transactions on Services Computing, vol. 15, no. 3, pp. 12341245, 2022.

10. P. Verma, Legal Validity of Electronic Signatures Under the Information Technology Act 2000, Indian Journal of Law and Technology, vol. 18, no. 2, pp. 89104, 2022.

11. O. Kovtun and O. Tulyakov, Prospects of Business Process Manage-ment Based on Chatbots, Problems and Perspectives in Management, vol. 22, no. 2, pp. 178189, 2024.

12. R. Patel and S. Mehta, WhatsApp-Based Order Management System for SMEs in the Indian Textile Industry, International Journal of Information Management, vol. 45, pp. 112121, 2019.

13. A. Rodrigues, B. Costa, and C. Ferreira, Technical Challenges in WhatsApp Business API Integration for Enterprise Applications, in Proc. International Conference on Web Engineering, 2022, pp. 234248.

14. A. Bhardwaj and V. Rathi, Comparative Analysis of Dynamic PDF Generation Approaches for Enterprise Document Automation, Interna-tional Journal of Document Analysis and Recognition, vol. 25, no. 1,

    pp. 4558, 2022.

15. K. Marciniak, P. Kowalski, and A. Nowak, Visual Field Mapping Interface for Legal Document Template Automation, in Proc. ACM Symposium on Document Engineering, 2021, pp. 110.