COGNIZANT CONSULTING
Helping organizations engage people and uncover insight from data to shape the products, services and experiences they offer
Learn More

Contact Us

THANKS FOR YOUR INTEREST IN COGNIZANT.

We'll be in touch soon!

x CLOSE

Refer back to this favorites tab during today's session for access to your selections.
Refer back to this favorites tab during today's session for access to your selections.x CLOSE

Perspectives

Unlocking the Value Trapped in Account Receivables Using Artificial Intelligence and Analytics (Part one of a two-part series)

2019-06-18


By combining automation, artificial intelligence (AI)-enabled machine learning (ML) and analytics, finance departments can bring much needed proactive, intelligent-driven change to account receivables management via dashboards, and a tailored approach to collections, follow-ups and reconciliation.

Although considered equivalent to cash, account receivables is cash that is locked by a counterparty. This inflates operational expenses and represents trapped value.

Many companies have established invoice processing teams whose primary role is to ensure collection of receivables with minimum losses. But these are manually driven tasks that involve establishing credit policies and collection strategies, crunching data from different enterprise resource planning (ERP) systems and drawing insights, as well as communicating with customers. Moreover, many of the resulting decisions and actions are then made based on judgment using inefficient and outdated techniques such as aging reports.

Delays in collection impact companies in three ways — they entail opportunity costs, require carrying costs and tend to increase debt due to adverse impact on the working capital requirements. Based on academic research, 13% to 29% of total assets of a company consists of account receivables. We estimate that about 17% of credit sales are stuck in account receivables based on average industry days sales outstanding (DSO).

Figure 1 shows the median DSO by industry. Even a 30-day reduction in the DSO has the potential to release 10% of the money stuck in credit sales.

Figure 1

Source: https://www.pwc.com/gx/en/working-capital-management-services/assets/pwc-working-capital-survey-2018-2019.pdf

The following is our estimate for a company with $100 million in annual credit sales. The annual cost of carrying receivables is typically $3.7 million (assuming a borrowing rate of 3.75% — derived as LIBOR + 1%). Figure 2 illustrates the value trapped in account receivables with 60-, 45-, 30- and 15-DSO.

Figure 2

Delays in the collection process have a direct impact on the working capital. Reduced working capital in turn drives the company towards raising debt to finance their everyday processes, thus impacting the company debt levels. So it is imperative for any competitive company to keep a stronghold on the account receivables.

Receivables management and challenges

Selling on credit is a requirement in most industries to stay competitive. However, it is marred with “pay late” and “pay over time” practices. This results in inordinate amount of value trapped in account receivables.

While most companies have a credit and collections policy/strategy in place to handle account receivables, 31% of companies do not. Large companies have separate teams to run this process. The collections team is typically responsible for invoice collection and the credit team is responsible for offering credit and providing approvals.

Broadly the invoice collections process involves the following:

  • Data assimilation

  • Tracking of receivables

  • Customer communication

  • Receipt reconciliation

Among the major issues in the receivables management and collection processes are disparate, non-intelligent IT systems, inaccuracies and inconsistencies of information across functions, lack of appropriate key performance indicators (KPIs), reactive processes and resource constraints.

  • Disparate and non-intelligent IT systems. Data pertaining to the receivable such as aging, counterparty credit score, payments, payment-related communication with counterparty, and material financing information is typically located in systems that do not communicate with each other. Therefore, all such information is collated manually by the user from disparate systems. In cases where relevant information is available under one platform or a report, the platform or report is not intelligent enough to draw insights to help the user make a proactive decision.

  • Inaccurate data and inconsistent information among collections, finance, treasury, operations and customers. Generally, restricted access to IT systems is provided to different user groups. For example, operators who raise invoices do not have access to the finance systems that contain details pertaining to the financing of the goods being sold. As a result, payment information is selectively duplicated across systems leading to redundancy and expanding the scope for data inconsistency.

  • Prioritizations and follow-up based on intuition and loosely defined KPIs. Worklist prioritization and collection tasks are carried out based on intuition, experience, and static parameters such as due date, invoice value or aging.

    Aging reports are used by many companies to track due payments. On the basis of the data in this report, the collections team draws insights, follows up with customers for payments and communicates the payment updates to internal functions such as treasury and finance. Also, the actions taken by the team generally follow a one-size-fits-all strategy, making this process less thorough and ineffective.

  • Reactive processes and issue management. Current business processes are reactive in nature and invoke action only in response to events that have occurred. For example: follow-ups occur only after an invoice becomes delinquent or shifts to a larger aging bucket. Prioritization, tracking, collection strategy or follow-up is driven by lagging factors such as the invoice value or shifts in the aging bucket — rather than proactive indicators such as trapped value, counterparty credit risk or payment probability. As a result, an inordinant amount of time is spent in attending to accounts with poor conversion ratios.

    Issues and disputes are escalated to relevant teams and higher authorities only when they are at the verge of occurring.

  • Limited staffing. Increasing volume of trade has resulted in an increase in the volume of invoices. A lack of staff to support such high volumes causes further inefficiencies.

Machine learning & analytics

The account receivables payment and collections strategy is more reactive than proactive. Since it is a highly rule-based and manual process, it also lacks the level of customization and rigor that technology-assisted processes have. With the advent of modern digital technologies, this process can be made technology driven.

Technology can be used to proactively manage account receivables and define collection policies by keeping the creditor and the debtor at the center of the strategy. Collection activities and payment from counterparties are influenced by micro factors such as number of delinquent invoices, exposure to a counterparty, invoice aging, working capital, debt levels, and are influenced by macro factors such as counterparty risk ratings, loan interest rates.

Artificial intelligence (AI) and analytics can be leveraged in this area to create proactive strategies that evolve and adapt as the macro and micro ecosystem changes (changes to internal aspects of the company such as number of delinquent invoices, exposure to a counterparty, invoice aging, working capital, loans, debt, etc.). It can assist the user in decision-making by:

  • Prioritizing worklist items. Invoices to be prioritized based on payment probability and maximum value realization.

  • Suggesting correspondence actions. When to send reminders, whom to send them to, and via which mode of communication.

Machine learning (ML) can retrieve and analyze historical customer data, retrieve insights on customer behavior and company cashflow, and learn from it. It can further make better decisions and act without human intervention.

Predictive analytics powered by a ML algorithm can be used to predict parameters such as expected payment dates on invoices or working capital at a point of time in the future. Prescriptive analytics can suggest actions such as incentives to be given to customers to expedite the payment on the receivables.

The digital solution

We recommend a four-pronged, new digital-empowered solution to overcome the challenges and, more importantly, arm companies’ account receivables management function with new capabilities to make the process more effective and result-oriented. The proposed solution includes development of ML-enabled decision-support dashboards, analytics-led capability to tailor customer-specific strategies, automation of follow-ups and reconciliation.

Equipped with these capabilities, companies can gear up to revitalize their account receivables management function and reap the resulting rewards in terms of contribution to their bottom lines. Read “A Digital Approach to Account Receivables (Part two of a two-part series)” for a deeper dive into our approach.

This article was written by Vinod Malpani, Pramit Basu and Naveen Krishnan from Cognizant Consulting’s Banking and Financial Services Practice.

To learn more, visit the AI & Analytics section of our website.

Related Thinking

Save this article to your folders


Save

PERSPECTIVES

How Machine Learning Can Lead the AI...

Advanced artificial intelligence in the form of machine learning holds the...

Save View

Save this article to your folders


Save

PERSPECTIVES

Three Steps to Propel Process...

North American financial services organizations are struggling to...

Save View

Save this article to your folders


Save

PERSPECTIVES

AI Inflection Point: Balancing Responsible...

To usher AI into the business mainstream, companies need to complement...

Save View
Unlocking the Value Trapped in Account Receivables Using Artificial Intelligence and Analytics (Part one of a two-part series)