Of the processes related to accounts receivable (AR), matching payments data to remittance data can be one of the most tedious and time-consuming, especially when it’s manual. But matching these data sets accurately is essential for businesses of all sizes, especially when it comes to efficient cash flow management.

When remittance matching is manual, it’s prone to errors, which cause payments to be delayed. As evidence, PYMNTS’ 2023 report, Automation Clears the Path to Getting Paid on Time, found that invoicing errors and discrepancies have caused payment disruptions for 45% of CFOs. Capgemini’s World Payments Report of 2023 says that collection and reconciliation processes in AR have caused problems for over 70% of corporate treasurers. 

Remittance processes are also expensive. Recent studies indicate that the average remittance costs 6.3% of the total amount paid. While many companies have moved from checks to electronic payments for greater efficiency, electronic payments introduce more complexity to remittances because the associated remittance information is sent separately from the payment. Moreover, electronic payments can come from a variety of payment channels — ACH, wire, real-time payments, or cards — and in different formats.

Fortunately, automation can dramatically improve the AR function, especially when enhanced by AI. Applying AI technology to remittance matching and processing makes cash flows more efficient, cuts costs, improves accuracy and reduces fraud, allowing finance professionals to focus on higher-value tasks. 

Automated AR

When AR is automated, technology is used to automatically extract and match payments data to remittance data. Using machine learning and analytics, automation solutions can generate invoices, send payment reminders, reconcile incoming payments and produce financial reports related to receivables, making AR processes more adaptable and scalable than ever before.

Even more important, payments come in faster. According to PYMNTS’ “B2B Payments Innovation Readiness Playbook,” firms that have automated their payment acceptance processes have an average days sales outstanding (DSO) of 40 days, while firms that have not automated AR have an average of 47 days. Plus, 87% of firms with automated AR say that their overall process speed has improved.

How it works

The process begins with the automatic creation of invoices using data extracted from a business’s financial systems. It then sends digital invoices directly to the customer or buyer.

As the resulting payments flow in, the solution matches invoices to payments and remittance data. Automated reminders can be sent to buyers/customers to alert them of imminent due dates or overdue payments. 

Automated AR solutions typically include reporting tools that help companies track the status of their receivables. Real-time analytics provide insights on customers’ payment behavior, anomalous activity and patterns, and more.

Advanced Automation Through AI

AI-powered solutions take these capabilities a step further. They learn an organization’s remittance processing rules and apply them consistently. They can automatically capture, classify and digitize remittance advices from files of all formats such as emails, PDFs, papers, handwritten notes, vendor portals, EDI, and lockbox and bank statements. The AI-powered solution matches remittances with open invoices and identifies unapplied payments, even if only a partial invoice number is available. To find patterns and flag any discrepancies between data sets, AI has machine learning algorithms that are trained to recognize and extract data correctly. 

Over time, the system “learns” from user interactions and historical data. Machine learning models become better at identifying patterns and matching accurately. The technology can detect errors such as missing or incorrect information, overpayments, short payments and inaccurate pricing. It will then mark the resulting exceptions for review and prompt resolution. If information is missing, some solutions have a digital assistant that goes directly to the customer to obtain the needed data.

When not using paper, approximately two-thirds of businesses send invoices via email as PDFs, as our recent guide to accounts payable automation notes. AI is particularly useful in these cases, because it learns to detect emails with remittance information by looking for keywords and attachments. It can identify and extract the necessary data even when the remittance is included in the body of the email rather than as a separate file.

AI solutions work optimally when deployed in a cloud system. Provider High Radius reports that cloud-based solutions can process and close more than 95% of invoices without any human intervention. These solutions can process data from different payers within a single entity and across customers. They learn and apply those lessons across accounts and companies. They eventually know where to look for information and how to classify it. Then they apply validation rules to ensure accuracy. If these solutions cannot find what they’re looking for, the user can train them to apply what they learn if that particular situation arises again.

Why AI?

The advanced capabilities of AI automation dramatically improve efficiency in remittance matching and reduce or nearly eliminate errors, making time-consuming, labor-intensive exception resolution a rare event. AI-driven solutions facilitate working capital optimization while also thwarting fraud, improving the customer experience, enhancing buyer-supplier relationships and more. 

Efficiency, accuracy and accelerated cash flow

AR automation rates gradually improve until exceptions requiring human intervention become few and far-between, making payment delays much rarer. If the AI-driven matching system is integrated with accounting and ERP systems, it can seamlessly update financial records and maintain them accurately.

With AI automation, remittance matching can achieve straight-through processing rates of up to 90%. As it learns, AI technology streamlines back-office operations, settlement and reconciliation. Deluxe recently reported that one of its corporate clients reduced the time it took to match remittances from as much as 12 hours to a mere 15 minutes. This time could drop to two minutes as the solution continues to learn. A study by BlueCreek Software revealed that a healthcare provider’s AR team decreased the 4.5 hours it spent daily on remittance matching by 75%. Now the company has automated nearly 90% of its incoming payments. 

In addition to expediting payment receipt and facilitating a robust supply of liquidity, AI-driven technology can generate accurate cash flow forecasts by linking to ERP systems’ general ledger transaction data, past bank statements and FP&A data. Multiple machine learning algorithms continuously analyze data and yield insights on potential AR and treasury scenarios. 

Fraud detection and prevention

AI’s fraud-fighting capabilities take prevention and detection to a new level. Machine learning and other AI-driven tools can analyze vast troves of data to detect duplicate or fraudulent invoices and payments as well as unusual spending and buying patterns — anomalies that humans might not notice or would need hours (or even days) to find and resolve. AI can even predict potentially fraudulent actors posing as customers and perform KYC activities. Having such tools protects not only financial security but also buyer-supplier relationships, which can go awry when fraud is suspected. 

As cybercriminals endlessly adapt to and overcome fraud prevention efforts, AI can adapt as well. It can answer questions about remittance fraud trends in certain regions or time periods. Using complex models, it can even predict fraud activity and schemes. Providing such information quickly or even in real time could prevent fraud altogether, before it is even attempted.

Improved customer experience

PYMNTS’ B2B payments innovation report revealed that 75% of firms with automated AR functions have significantly improved their customer experience. If automated solutions include generative AI, virtual remittance assistants can provide 24/7 customer support. These chatbots can be trained on extensive data sets that equip them to instantly supply answers that otherwise might take days to provide. Customers/buyers receive accurate, up-to-date information faster than ever before. 

Better buyer-supplier relationships

AI-driven remittance processing can improve buyer-supplier relationships by virtually eliminating errors in a formerly error-prone process. Errors not only delay payments but can also lead to disputes that may strain these relationships. With fewer exceptions to resolve, employees can avoid badgering customers to correct their mistakes — and save time and costs in the process.

Risk management and compliance

Some of the more advanced AI-powered solutions can manage risk by assessing a customer’s credit-worthiness using integrated data from credit rating agencies or machine learning algorithms. Most solutions provide real-time data to ensure accurate records and document transactions in compliant formats. They can maintain those records in accordance with applicable regulations, accounting rules and reporting requirements to make compliance and audit preparation much easier.

Next Steps to AI-Driven Automation

Successful deployment of AI in remittance matching — and attaining high, straight-through processing rates — hinges on factors like the data received from customers, AR systems and access to data from the ERP system. Connecting to these data sets allows AI to match every piece against the others. 

When a business decides to adopt AI-driven remittance matching, it should first determine which pieces of data need to be matched, such as pricing and invoice numbers. When searching for a solution, companies should note that available products and services have a wide range of features, so making a list of specific needs and processes is key. Once the solution has been integrated with AR systems, staff should be trained to make the most of its capabilities effectively and efficiently. Then they can focus on more challenging, meaningful tasks than the tedium of remittance matching. 

AI technology will only improve over time as it learns and is applied to more advanced models. Even now, using AI for remittance matching and processing can make cash flow into a business quickly, building ample liquidity while slashing costs, ensuring accuracy, detecting and preventing fraud — and generating reporting and insights like never before.