How to track outstanding payments using reconciliation

Outstanding payments create cash flow uncertainty, manual work, and friction with customers and partners. Reconciliation is the operational process that surfaces these unpaid or mismatched items by comparing internal records to external statements.

This article walks finance teams through a practical reconciliation-driven workflow to find, prioritize, and close outstanding payments. It covers the data you need, matching logic, AI-assisted exceptions, and the follow-up actions that reduce days sales outstanding.

Use the steps below to turn unreconciled transactions into a controllable list of actions you can automate and resolve consistently.

Why this topic matters

For controllers and finance managers, outstanding payments are both an operational headache and a business risk. Unreconciled transactions can hide lost revenue, duplicate collections, disputed refunds, or reconciliation timing differences that distort cash forecasts.

A structured reconciliation process gives visibility into where money is expected but not received, who to contact, and which items need investigation versus those that are true timing differences. That visibility reduces manual audit work and improves cash recovery rates.

Core components

Successful tracking of outstanding payments depends on three core components: accurate side-by-side data, a robust matching engine, and a clear exception workflow.

Side A and Side B: what to collect

• Side A (internal): invoice records, ERP sales ledger, accounts receivable extract, order reports, or general ledger entries. Include invoice ID, customer ID, date, gross/net amounts, and status.
• Side B (external): bank statements, payment gateway payouts, marketplace settlements, or partner remittance files. Include transaction reference, posting date, and settled amount.
• Supporting data: customer master, fee schedules, returns or refunds file, and mapping tables to enrich matches and explain amount differences.

Collecting clean, well-structured files on both sides is the foundation of reliable accounts receivable reconciliation and subsequent payment tracking.

Data standardization and derived columns

Before matching, normalize dates, amounts, and identifiers. Common cleanups include trimming whitespace, converting currency formats, and standardizing order or invoice IDs.

Derived columns are particularly useful when the source files don’t contain a ready-made amount or identifier. Examples:

• Net amount after fees: calculated from gross amount minus fee columns.
• Matching key: concatenate customer ID and order ID to improve deterministic matching.

Deriving the right columns improves matching accuracy and reduces manual review.

Matching logic and AI-assisted matching

A reliable reconciliation engine uses layered logic:

• Deterministic rule-based matching: exact identifier or reference matches, date + amount matches, and configured one-to-many/grouped matches.
• Relaxed and similarity rules: contains, similar text, or subset matches when identifiers are inconsistent.
• AI-assisted matching: analyzes remaining exceptions where references are unstructured, identifiers differ, or group/net matching is needed.

The engine should classify results as fully matched, partially matched (identifier matches but amounts differ), unmatched, or skipped (invalid or incomplete records). This classification creates a prioritized queue of outstanding payments.

Practical implementation steps

Follow these step-by-step actions to start tracking outstanding payments using reconciliation.

Step 1: Prepare and upload files

1. Export the required Side A and Side B reports into CSV/XLSX formats.
2. Ensure each file uses a consistent header row and includes a date, amount, and at least one reference or identifier where possible.
3. Upload files to your reconciliation platform and confirm they pass the format checks.

Tip: If files vary across periods, create a reusable template so future uploads don't fail due to column mismatches.

Step 2: Configure columns and derived fields

• Select the header row, date column, amount column, and identifier columns on both sides.
• Add supporting files such as fee schedules or return reports to adjust amounts.
• Create derived columns where needed, for example to calculate net settled amounts or to normalize identifiers.

These configurations ensure the matching engine has the right inputs to reconcile reliably.

Step 3: Run rule-based matching

• Start with deterministic rules: exact ID equals ID, amount equals amount, and date windows where appropriate.
• Allow one-to-many and many-to-one configurations for scenarios like aggregated settlements versus detailed invoices.
• Review the fully matched set to confirm high-confidence matches are complete.

This step removes the low-effort items from the queue, leaving the real exceptions.

Step 4: Review AI suggestions and exceptions

• Let the AI layer propose matches for records that lack exact identifiers or have inconsistent references.
• Inspect partially matched items carefully; these often indicate fees, partial refunds, or short-payments.
• Tag items that require further investigation, such as suspected duplicate payments or potential chargebacks.

AI suggestions accelerate discovery but should be reviewed before any accounting adjustments.

Step 5: Manual reconciliation and follow-up

• For true unmatched invoices, create a prioritized follow-up list. Use aging buckets to prioritize older receivables.
• Record manual matches when evidence from emails, remittance advice, or customer confirmations resolves an item.
• Export audit-ready reconciliation reports and attach supporting evidence for each resolved exception.

Automation options: schedule uploads, run reconciliations automatically, and push reports to accounting systems so follow-ups become part of a regular AR cadence.

Common mistakes to avoid

• Uploading inconsistent file formats without standardizing headers, which causes rejected files or skipped records.
• Relying only on date-and-amount matching when reliable identifiers exist; this can create false positives.
• Ignoring partially matched items; they often hide fee discrepancies or short payments that affect cash reconciliation.
• Overdependence on automated matches without human review for high-value or disputed items.
• Not capturing root causes: resolving items without logging why they were outstanding reduces process improvement opportunities.

Key Takeaways

• Use clean Side A and Side B data plus supporting files to improve matching accuracy.
• Start with deterministic rules, then apply AI for unstructured or grouped exceptions.
• Prioritize follow-up using aging buckets and partially matched items to protect cash flow.
• Record manual matches and retain evidence to produce audit-ready reports.
• Automate routine uploads and reconciliation runs to reduce manual effort and ensure consistent tracking.

Conclusion

A repeatable reconciliation process turns unknowns into a manageable pipeline of outstanding payments that finance teams can prioritize and close. By combining data standardization, deterministic matching, and an AI layer for exceptions, you reduce manual work and improve cash visibility.

Implementing outstanding payments reconciliation as described above will make your accounts receivable reconciliation more efficient and auditable. The next step is to try a reconciliation platform that supports derived columns, mixed matching rules, manual matching, and audit-ready exports.

Start your 14-day free trial with Cointab. No credit card required. 14-day free trial.