Degree

Bachelor of Science (Computer Science)

Department

Department of Computer Science

School

School of Mathematics and Computer Science (SMCS)

Advisor

Ms. Tasbiha Fatima, Lecturer, Department of Computer Science

Keywords

Payment Reconciliation, FinTech, Process Automation, Transaction Processing

Abstract

Pakistan’s digital payment ecosystem is growing rapidly, placing increasing demands on payment aggregators to maintain accurate and scalable transaction reconciliation. The challenge for PayPro, who serves as a payment aggregator for gateway companies such as Meezan Bank (MBL) and 1Link, was a critical operational hurdle: a completely manual process of reconciling payments that required 4 to 6 hours per day for their finance team. The process was prone to error, difficult to scale and posed a compliance risk by relying on spreadsheet comparisons, manual file transfers, and institutional memory. To address this challenge, this project designed, engineered, and deployed an automated and serverless reconciliation system. The development approach chosen was to create data parsing algorithms in Python and add intelligent discrepancy-matching logic, which would be distributed by Zapier’s workflows and cloud storage triggers, without requiring any specific server infrastructure. The proposed solution involves legacy-specific cleaning modules to manage unstructured and inconsistent data formats, an intelligent reconciliation engine that can reconcile data in both directions, and a settlement summary generator that can be automated and enriched with the bank code for direct submission to MBL. The system was able to automate a manual six hour process into a pipeline that runs in less than half an hour. Results show an 82% decrease in MBL reconciliation time, an 85% decrease in 1Link reconciliation time and eliminate data entry and bank code mapping errors 100%. A retail settlement cycle was curtailed by one single business day while a detailed audit trail was developed to meet regulatory requirements of State Bank of Pakistan. Modular and gatewayagnostic architecture provide PayPro with a scalable backbone for future growth.Pakistan’s digital payment ecosystem is growing rapidly, placing increasing demands on payment aggregators to maintain accurate and scalable transaction reconciliation. The challenge for PayPro, who serves as a payment aggregator for gateway companies such as Meezan Bank (MBL) and 1Link, was a critical operational hurdle: a completely manual process of reconciling payments that required 4 to 6 hours per day for their finance team. The process was prone to error, difficult to scale and posed a compliance risk by relying on spreadsheet comparisons, manual file transfers, and institutional memory. To address this challenge, this project designed, engineered, and deployed an automated and serverless reconciliation system. The development approach chosen was to create data parsing algorithms in Python and add intelligent discrepancy-matching logic, which would be distributed by Zapier’s workflows and cloud storage triggers, without requiring any specific server infrastructure. The proposed solution involves legacy-specific cleaning modules to manage unstructured and inconsistent data formats, an intelligent reconciliation engine that can reconcile data in both directions, and a settlement summary generator that can be automated and enriched with the bank code for direct submission to MBL. The system was able to automate a manual six hour process into a pipeline that runs in less than half an hour. Results show an 82% decrease in MBL reconciliation time, an 85% decrease in 1Link reconciliation time and eliminate data entry and bank code mapping errors 100%. A retail settlement cycle was curtailed by one single business day while a detailed audit trail was developed to meet regulatory requirements of State Bank of Pakistan. Modular and gatewayagnostic architecture provide PayPro with a scalable backbone for future growth.Pakistan’s digital payment ecosystem is growing rapidly, placing increasing demands on payment aggregators to maintain accurate and scalable transaction reconciliation. The challenge for PayPro, who serves as a payment aggregator for gateway companies such as Meezan Bank (MBL) and 1Link, was a critical operational hurdle: a completely manual process of reconciling payments that required 4 to 6 hours per day for their finance team. The process was prone to error, difficult to scale and posed a compliance risk by relying on spreadsheet comparisons, manual file transfers, and institutional memory. To address this challenge, this project designed, engineered, and deployed an automated and serverless reconciliation system. The development approach chosen was to create data parsing algorithms in Python and add intelligent discrepancy-matching logic, which would be distributed by Zapier’s workflows and cloud storage triggers, without requiring any specific server infrastructure. The proposed solution involves legacy-specific cleaning modules to manage unstructured and inconsistent data formats, an intelligent reconciliation engine that can reconcile data in both directions, and a settlement summary generator that can be automated and enriched with the bank code for direct submission to MBL. The system was able to automate a manual six hour process into a pipeline that runs in less than half an hour. Results show an 82% decrease in MBL reconciliation time, an 85% decrease in 1Link reconciliation time and eliminate data entry and bank code mapping errors 100%. A retail settlement cycle was curtailed by one single business day while a detailed audit trail was developed to meet regulatory requirements of State Bank of Pakistan. Modular and gatewayagnostic architecture provide PayPro with a scalable backbone for future growth.

Tools and Technologies Used

Python, pandas, regex, Zapier, Google Drive API, Streamlit, Gmail (automated notifications), Microsoft Excel (openpyxl/formula-protected outputs), CSV, Google Drive (file-based trigger and storage)

Methodology

We developed our system using an Agile-Scrum sprintbased approach, starting from gathering requirements and data analysis for gateway-specific implementation and deployment. The architecture follows a stateless three-layer pipeline: a file-based ingestion layer using Google Drive folder monitoring, a Python processing engine with custom per-gateway parsers, ID normalization, matching logic, and settlement aggregation, and an output and delivery layer producing CSV reports, Excel summaries, automated email notifications, and a dated folder audit trail. Key design principles include stateless processing with no persistent database dependency, sequential execution for dependency safety, per-step error isolation, and a gateway-agnostic output format. The originally proposed database-centric approach using SQL Server and Power Automate was replaced with this serverless file-based architecture after discovering that MBL provides no programmatic API access and 1Link's FTP uses undocumented authentication, making file-based processing the only practical path to full automation.

Document Type

Restricted Access

Submission Type

BSCS Final Year Project

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS