Get a personalized demo of Stuut and see how it can help with AR automation.
Cash application exceptions are some of the most time-consuming parts of AR because resolving them requires gathering evidence across multiple systems, customer portals, and inboxes before a single dollar can post. Partial payments with no remittance, bulk wires covering multiple invoices, and deductions with no reason code all create the same bottleneck: cash you've already received sitting in a suspense account while your AR aging report shows those invoices as open and your DSO climbs.
This guide breaks down the exact workflows for handling short pays, unapplied cash, and deductions, and shows how AI agents automate the evidence-gathering so your team can focus on the decisions only a human can make.
Unapplied cash is money your company has received and deposited into your bank account but has not yet matched to a specific open invoice or customer account in your AR ledger. Until you match it, that cash sits in a suspense account, inflating your bank balance while your AR aging report still shows those invoices as open.
Manual invoice processing delays across AR labor, portal access time, and follow-up drag on Days Sales Outstanding (DSO) because cash you've collected doesn't reduce your reported receivables until it posts to the correct invoice.
A short pay is any partial payment of an invoice. The cause matters more than the amount, because the resolution path differs completely depending on why the customer paid less.
Short pay vs. deduction at a glance:
Deductions are intentional reductions the customer applies when paying, and may be valid (backed by a contract, trade promotion agreement, or payment term) or invalid (no contract backing). Unexpected short pays are discrepancies with no stated reason that require investigation before you know whether the customer owes the difference.
Remittance advice is the document a customer sends alongside a payment that identifies which invoices the payment covers and in what amounts. When remittance is missing or incomplete, you can't post the payment without manual research, and that creates a suspense account backlog that worsens every day you don't resolve it.
A suspense account is a temporary holding account for transactions where there's uncertainty about the correct posting destination. The goal is always to clear suspense accounts before month-end close, because unresolved balances distort your AR aging report and your DSO calculation.
Three-way matching is the process of verifying a payment by cross-referencing three documents: the purchase order (PO), the receiving report, and the supplier invoice. All three must agree on quantities, prices, and delivery terms before a payment can post cleanly.
The three documents are:
Bulk deposits make 3-way matching significantly harder. When a single Stripe deposit covers many separate customer payments, you need to break that deposit into individual sub-payments and match each one to its originating invoice before you can post any of them.
Deductions extend well beyond trade promotions. Common categories in manufacturing and distribution include freight or shipping charges, damaged goods on delivery, unauthorized early-pay discounts, vendor compliance fines from large retailers, and return chargebacks. Each category requires different backup documentation and has a different resolution path. The reason code attached to a deduction determines both who owns the resolution and what evidence you need to validate or reject the claim.
Many cash application exceptions are preventable. Common upstream causes include invoices sent to the wrong AP contact, missing PO numbers on the invoice, billing emails that bounce because a customer's AP team changed, and payment terms that don't match what the customer's system recorded at order time. Auditing your invoice template for these fields and confirming AP contacts during customer onboarding reduces your incoming exception volume before it hits the matching queue.
The manual workflow for resolving exceptions follows a consistent sequence regardless of exception type. Getting this sequence right reduces resolution time and protects your ability to recover invalid deductions before filing windows close.
Pull your exception report from the ERP and sort it by suspense account balance. Your goal is a prioritized list showing each unmatched payment, its amount, the customer account it came from if identifiable, and how many days it has been sitting unmatched. Most ERP systems surface this through an unapplied payments or unallocated receipts view in the receivables module.
Sort exceptions by three criteria in this order:
This prioritization ensures your team works the exceptions with the highest business impact first, rather than processing them in arrival order.
For a short pay caused by a billing error, a direct email to the AP contact asking them to confirm whether the difference relates to a specific deduction or whether they need a corrected invoice typically resolves the issue quickly. For a short pay where the customer took an unauthorized early-pay discount, a brief phone call works better because the conversation requires confirming payment terms and agreeing on a repayment or credit memo approach.
Once you've confirmed whether a deduction is valid or invalid, the resolution follows one of three paths:
Standardizing how your team classifies exceptions saves resolution time and creates the reporting data you need to identify upstream problems in pricing, shipping, or order management.
ERP systems like SAP, Oracle, and NetSuite let you define custom reason codes for deductions. Most AR teams standardize around a core set that covers the most frequent scenarios in their industry:
Using consistent codes matters because inconsistent reason code usage creates reconciliation errors at month-end and makes it impossible to report on deduction trends by root cause.
Valid deductions trace back to a specific contractual provision: a trade promotion agreement, a freight terms clause in the master service agreement, or an early-pay discount written into the payment terms. Invalid deductions have no contract backing.
Your validation checklist for any deduction:
If all four checks pass, apply the credit memo. If any fail, flag it as invalid and start the recovery process.
Balances below your materiality threshold cost more in analyst time to resolve than their recovery is worth. Document your write-off threshold in your collections policy, apply a de minimis reason code, and close those invoices monthly rather than letting them age and distort your aging buckets. Most AR teams set this threshold based on average invoice value and the cost of a single resolution cycle.
Routing exceptions to the right internal owner prevents them from sitting in the AR queue when resolution requires action from another department. A pricing error belongs to the sales rep who quoted the wrong rate. A damaged goods deduction belongs to the shipping or operations team that can pull the delivery documentation. A vendor compliance fine belongs to supply chain. Build an escalation routing map in your ERP or CRM that automatically assigns exception types to the correct owner.
Exception handling is an evidence problem. The faster you gather the right documents, the faster you post cash and close invoices.
Remittance advice arrives through several channels: emailed PDF attachments, bank lockbox files in EDI format, and data attached directly to ACH or wire transactions. Parsing each format manually takes time and requires different export steps depending on your bank and ERP. For lockbox remittance, pull the daily file from your bank portal, import it into the ERP, and let the system attempt automatic matching before routing residual exceptions to your team.
Logging into customer portals like Ariba, Coupa, or Tungsten to retrieve backup documentation for a deduction can be a time-consuming part of the AR workflow. Each portal has a different interface, different download formats, and different filing deadline structures. Retrieving that documentation still means logging into each customer's portal individually, since remittance and backup data live on their system, not yours, which means navigating dynamic web interfaces and manually linking documentation to the right dispute case.
For damaged goods and short shipment deductions, your strongest evidence is the signed proof of delivery (POD) from your carrier and the packing slip the customer signed at receipt. Pull both before you contact the customer, because the conversation goes faster when you can reference specific document numbers and dates rather than asking the customer to locate their own records.
AI handles exception triage faster than human analysts because it processes remittance data and ERP records simultaneously rather than checking each source sequentially. The critical distinction is that AI handles the evidence-gathering and routine matching, while human analysts handle the negotiations and judgment calls on complex disputes.
Stuut's proprietary 3-way matching algorithm achieves a 95%+ automated cash application match rate by learning remittance patterns specific to each customer over time. The system stores metadata that most ERPs never capture, including bank transaction identifiers and originating company numbers, and uses that metadata to match future payments from the same source instantly rather than routing them to a manual queue.
For bulk deposits, the algorithm breaks the deposit into individual sub-payments, matches each one to its originating invoice, and posts the cash application entries to your AR subledger in real time. This eliminates the payment matching bottleneck that delays month-end close.
When Stuut identifies a deduction it cannot match automatically, it categorizes the exception by reason code, attaches the supporting documentation retrieved from the bank file or remittance data, and creates a dispute case in your workflow system (SAP, Salesforce, or equivalent). That replaces the manual triage per exception that currently consumes analyst time at scale. Tools like Billtrust centralize dispute handling in a management portal where your team works cases to resolution, whereas Stuut is designed to execute the credit memo and closure autonomously, so your team reviews the outcome rather than performing each step.
Stuut's self-learning intelligence remembers every exception resolution and applies that knowledge forward. If a customer always takes the same early-pay discount on the same day each month, the system learns that pattern and applies the correct credit memo automatically on future payments, without a manual rule update. Stuut's learning compounds across a portfolio over time. PerkinElmer, for example, reduced overdue invoices from 50% to 15% in one year using Stuut, collecting $300M and automating outreach to 80% of tail customers.
Stuut continuously monitors all open invoices, payment activity, and customer communication in real time. When it detects an anomaly, such as a payment pattern that suggests a deduction is incoming or an unresponsive customer with an invoice approaching 60 days, it alerts your team proactively so you can act before the exception ages. This shifts exception management from reactive (discovering problems during the daily aging review) to proactive (resolving them before they distort month-end close).
Leaving unapplied cash in suspense through month-end close creates two problems: your AR aging report overstates open receivables, and your DSO calculation inflates because cash you've already collected hasn't reduced your reported balance. Suspense accounts are temporary holding spots, not permanent homes for unmatched transactions. Clear them before close.
Applying the wrong reason code to a deduction creates downstream reporting problems that compound over time. If you code a pricing error as a short shipment, your operations team never sees the pattern that a specific sales rep is systematically quoting the wrong price. Poor reason code standardization forces teams into spreadsheet workarounds that introduce additional coding errors and make it impossible to spot operational trends.
Tracking how long each exception type takes to resolve reveals operational problems that originate outside AR. If pricing error deductions consistently take longer to close, the delay usually means the sales team is slow to confirm correct pricing. If damaged goods deductions lag, operations may not be pulling PODs quickly enough to meet retailer filing deadlines. This diagnostic visibility is one of the clearest arguments for standardized reason codes and resolution time tracking.
Use this checklist in the five business days before close to clear your exception queue:
Every exception that clears before month-end close converts unmatched cash into posted receivables and removes it from your suspense account backlog. That directly reduces your reported DSO because cash you've already collected stops inflating your open AR balance. How quickly your team works through this sequence determines how much working capital stays trapped versus available to fund operations.
If you're evaluating AI to handle exception resolution automatically, book a demo to see how the 3-way matching algorithm handles bulk deposits, remittance parsing, and credit memo creation without manual intervention.
A short pay is any partial payment of an invoice, whether the reason is valid or not. A deduction is a specific type of short pay where the customer intentionally pays less than the full invoice amount, typically due to a perceived discrepancy or dispute, such as damaged goods, a pricing error, a promotional allowance, or an early-pay discount.
Manual triage on a single exception takes meaningful analyst time, and complex deductions requiring portal documentation can take significantly longer to fully resolve. AI-automated triage reduces initial categorization to seconds, with resolution time depending only on how quickly the customer or internal stakeholder responds.
AR analysts own the initial triage and evidence gathering. Resolution responsibility splits by type: pricing errors route to sales, damaged goods and short shipments route to operations or shipping, and unauthorized discounts stay with the AR analyst to request repayment directly from the customer.
Stuut's overall automated cash application match rate is 95%+, covering exact matches, partial payments, overpayments, and bulk deposits. That figure reflects the full payment portfolio, not just exceptions. Complex multi-entity disputes and cases requiring customer negotiation still route to human analysts for resolution.
Unapplied cash: Money received and deposited but not yet matched to a specific open invoice in the AR ledger, held temporarily in a suspense account until matched.
3-way matching: The process of verifying a payment by cross-referencing three documents: the purchase order, the receiving report, and the supplier invoice, confirming all three agree before posting cash.
Remittance advice: A document sent by a customer alongside a payment that identifies which invoices the payment covers and in what amounts, enabling accurate cash application.
Suspense account: A temporary holding account in the general ledger used to store transactions where the correct posting destination is uncertain, cleared once the matching information is confirmed.
