In accounting terms, working capital includes current assets (short-term items such as cash, money due from customers and inventory) and current liabilities (typically payments due to suppliers and loan amounts that must be repaid within one year). Working capital management is a prime function of the finance organization, designed to balance often-conflicting objectives related to revenue, liquidity, risk and profitability. The basics of working capital management date back to ancient clay tablets, but today’s technology can enable organizations to handle these basic management tasks more efficiently and effectively than ever.
An organization must manage its current accounts to ensure sufficient liquidity to meet its immediate operating needs without stifling sales opportunities or putting its future financial condition at risk. The process is dynamic and requires ongoing forecasts to ensure that the components of working capital remain well-balanced over time.
Primary financial considerations in working capital management are costs, risk minimization, liquidity and efficiency (notably inventory optimization). There are also impacts on operations to consider. Timely payments to suppliers foster a positive relationship that reduces friction and can confer advantages, while defter handling of late-paying customers can help these relationships. Providing favorable credit terms helps foster sales, but too lax an approach risks losses from non-payment.
Technology can significantly enhance accounting staff productivity through automation. Artificial intelligence using machine learning will be an essential technology underpinning working capital management. Use cases that are either available or will be within the next three years include:
Working capital management lends itself to AI because most considerations are easily quantified, there is sufficient historical data to train models and a considerable percentage of actions are mechanical and repetitive.
AI is already at work in processing the paperwork around accounts payable and accounts receivable, albeit at a relatively low level of penetration. A wide variety of software vendors have made it possible to automate document ingestion by “reading” an invoice or purchase order document, for example, regardless of whether it’s physical paper, an electronic document (such as a PDF) or information in an email. Because the various layouts and typography of these documents are so well established, the systems can achieve initial high levels of accuracy and require limited training for a given organization. And trained systems are more likely to be accurate than humans in handling data entry.
Task supervision enables a system to spot a potential duplicate invoice or items not in line with experience, such as the quantity of a specific product ordered by a customer. These issues can be flagged for immediate attention following a consistent process. A system can be programmed to spot a lack of activity from a regular customer, which might indicate they are buying from another vendor. Or it can flag a customer who routinely pays on time but is now behind, which might indicate an issue to be resolved. Addressing that issue before sending a snarky dunning email supports customer satisfaction.
Particularly for smaller organizations, forecasting and managing cash positions isn’t complicated. Complications arise when there are cash accounts with multiple institutions, especially when these accounts are in multiple tax jurisdictions where cash transfers can trigger tax liabilities or default provisions from lenders. Dedicated treasury management software facilitates granular cash flow forecasting as well as managing cash balances, processing payments and providing dashboards and analytics that provide situational awareness and insight. Over the next five years, treasury management software will increasingly apply AI to generate detailed predictive analytic models that improve forecast accuracy and boost productivity by automating a substantial amount of the purely mechanical elements of treasury management.
One of the benefits of AI-assisted working capital process management is its ability to manage by exception, automating obvious steps in a process. When there is ambiguity, the system presents a set of choices, potentially with guidance to help make the best decision. An advantage of applying machine learning to processes is that as these issues are resolved manually, this information is incorporated into the system so more can be performed automatically.
Using AI-assisted software for working capital processes increases accounting staff productivity by automating process steps and ensuring the timely completion of handoffs, reviews and approvals. Accelerated processing also facilitates business operations. On the receivables side, for example, having the ability to credit a customer’s account with a payment makes it possible for the customer to buy more sooner if the company is at or near its credit limit. On the payables side, AI-assisted systems enable businesses to more readily maximize early payment discounts, cutting the cost of sales and achieving a higher return on cash balances compared to earnings on short-term deposits.
Admittedly, working capital management is not the most exciting topic, and the application of AI to handle these tasks is not an advanced technology use case that sets the heart racing. Nonetheless, eliminating a myriad of small frictions in business processes and increasing the productivity of those involved in mechanical, repetitive tasks is likely to have a notable, positive impact on an organization’s profitability. I urge finance executives to use AI extensively in working capital systems and processes as soon as possible since these systems aid profitability and enhance buyer and supplier relationships.
Regards,
Robert Kugel