Summary
The client needed daily three-way reconciliation of payment data across multiple systems. The existing process meant manually collecting payment information from two internal applications and a banking portal, then reconciling and preparing a report by hand.
We automated extraction, reconciliation and reporting. The robot now performs the full reconciliation across all three sources and shares a consolidated report, reducing daily processing time from 45 minutes to 20 minutes while improving accuracy.
The client
The client operates in a transaction-intensive environment where payments arrive through multiple channels, including online transactions and cash. Payment data lives across two internal applications, while bank confirmations are accessed through a banking portal. Finance and operations own daily reconciliation accuracy for financial integrity and compliance.
The problem
Reconciliation was fully manual and repeated every single day.
- 01Log into multiple internal systems to extract payment details
- 02Access the banking portal to download settlement information
- 03Compare data across three sources by hand
- 04Identify mismatches and prepare the reconciliation report
What it was costing
- 45 minutes of manual effort every day
- Increased risk of reconciliation error
- Delays in identifying discrepancies
- Limited audit trail and inconsistent report formats
Objectives
Goals
- Reduce manual effort in daily reconciliation
- Improve accuracy and consistency of results
- Speed up discrepancy identification
- Standardise reconciliation reporting
- Ensure reliable audit trails
Success criteria
- Daily processing time
- Error reduction
- Reconciliation completeness
What we built
One scheduled process that gathers from all three sources, matches on the fields that matter, and produces the same report in the same format every day.
Internal application 1
Payment data for online transactions extracted directly from the source system.
Internal application 2
Cash payment records extracted and normalised into the same shape.
Banking portal
The robot logs into the portal and pulls settlement data for the period.
Three-way match
Transactions matched across all sources on reference, date and amount.
Structured report
Matched records and unmatched or mismatched transactions separated, then emailed to users automatically.
Delivery
Phased delivery — discovery through post-go-live support
- Process discovery and reconciliation logic definition
- Extraction design across three sources
- Bot development and unit testing
- Validation with finance users
- UAT and production deployment
- Post-go-live support
Challenges and how we handled them
Stakeholders
Results
| Measure | Before | After |
|---|---|---|
| Daily processing time | 45 minutes | 20 minutes |
| Data extraction | Manual, three logins | Automated across all sources |
| Matching | Manual comparison | Rules-based on reference, date, amount |
| Report format | Inconsistent | Standardised and emailed |
| Audit trail | Limited | Detailed logs per run |
Quantitative
- Daily reconciliation time reduced from 45 minutes to 20 minutes
- Significant reduction in manual effort for finance teams
- Faster detection of discrepancies
- Improved SLA adherence
Qualitative
- Increased confidence in financial data accuracy
- Standardised, audit-ready reconciliation reports
- Repetitive work removed from the finance team's day
Business value delivered
Stronger financial controls
Reconciliation happens consistently every day without depending on anyone being available to do it.
Earlier visibility on discrepancies
Mismatches surface the same morning rather than at period end.
Extensible framework
Additional payment sources and reconciliation scenarios drop into the same structure.