Corporate Card Reconciliation Platform

The Challenge

A regional financial services group with 2,000+ corporate cardholders across 8 entities faced a monthly reconciliation crisis. Every month, the finance team spent 5+ FTE-weeks matching card statements against expense reports, chasing missing receipts, and investigating discrepancies.

The pain points:

• Card statements from 3 different issuers with inconsistent transaction formats
• Expense data split across two expense management systems (legacy and current)
• Multi-currency transactions with exchange rate timing differences
• Missing receipt rate of 15% requiring manual follow-up with cardholders
• Reconciliation delayed month-end close by 4–6 days

Our Approach

Discovery (2 weeks)

We mapped the end-to-end reconciliation workflow and quantified the manual effort at each step. The biggest surprise: 40% of the finance team's time was spent not on matching, but on chasing documentation and formatting data for comparison.

Solution Design

Rather than replacing the expense management systems, we built an intelligence layer that connected them:

Multi-source ingestion — Automated import from all 3 card issuers and both expense systems, normalising formats into a common transaction schema.

Configurable fuzzy matching — Matching rules that account for vendor name variations, date windows (up to 5 business days), amount tolerances (for FX rounding), and split transaction logic. Rules are different per entity to reflect local policies.

Exception classification and routing — Unmatched transactions are classified by type (missing receipt, amount mismatch, unknown vendor, policy violation) and routed to the right person with context — not dropped into a generic queue.

Learning engine — When a finance analyst resolves a match manually, the system captures the mapping. Vendor name aliases, recurring transaction patterns, and entity-specific rules accumulate over time.

Compliance dashboard — Real-time visibility into policy violations, unmatched transactions, and reconciliation status per entity.

Deployment (10 weeks)

Rolled out in 3 waves: pilot entity, high-volume entities, remaining entities. Each wave included training for finance teams and calibration of matching rules.

The Result

After full deployment across all 8 entities:

• 96% of transactions matched automatically without manual intervention
• Manual reconciliation effort dropped by 80%
• Month-end close shortened by 4 days
• Policy violations detected in real-time instead of 30 days after the fact
• Finance team shifted from data entry to exception analysis and policy improvement