FIS Modern Banking Platform
Advance your bank with a modern core platform.
Artificial Intelligence: Driving Cash Flow Improvements Across the Credit-to-cash Cycle
January 23, 2019
There is a lot of talk about Artificial Intelligence (AI) these days, especially in the collections arena. But what constitutes AI and is it really beneficial? What about teams that are highly functioning without using AI? The short answer is: Yes, it is beneficial, even for those teams that are performing well. Everyone that is selling and leveraging credit terms is capable of improving results.
Let’s start with a simple understanding of what constitutes AI. AI is the machine’s ability to learn and adapt without human intervention to continue to improve and obtain optimal results. This is a fairly vague statement, but it contains some key elements for defining AI. First, the machine’s ability to learn indicates that the machine can monitor some key data, action or process and understand what makes it good or bad. Second, the machine can take action without a human telling it to take that action. Now, this can be a little scary for people. Trusting that a machine is automatically taking the appropriate action. Lastly, the machine is continuously improving. This statement indicates that there isn’t one correct answer. Things change over time and so the AI engine must adapt to continue to achieve optimal results.
Given that AI is a relatively new approach to solving an old problem of an ever-changing environment, people are understandably hesitant to trust the machine making decisions on its own. Of course, movies are making matters worse, by showing the worst-case scenarios of AI running amuck. The reality is that AI can monitor more data than any individual or team of individuals can, to make accurate decisions within seconds.
Leveraging AI across the full credit-to-cash cycle affords greater combined benefits than simply applying it in one or even a few areas. For simplicity’s sake, let’s look at each area individually. Ask any credit manager how they view their world, they will tell you it is a combination of science and art. Loosely translated, there are so many potential data elements involved in assessing appropriate credit risk, one standard formula just does not provide enough information to make a sound decision. They may use a basic set of standard scores to give them a level of comfort, but they cannot possibly review all data elements for each customer to truly mitigate risk.
This is where AI can assist. It automatically pulls all of the necessary data. Think how much time is wasted by team members simply pulling the data to be reviewed. So much so that some companies have looked to “outsource” the pulling of the data to low-cost countries or operating units. The issue with this approach, however, is that delays are introduced in the process by adding in more handoffs and potential failure points. AI is then able to assess the internal and external data sources within seconds to determine the level of risk and automatically assign a credit line or move the request into a workflow for additional review. This provides more information for making risk decisions, speeds up the review process and frees resources to focus on more value-added tasks.
Believe it or not, AI has been used in the collections area for many years. It just wasn’t marketed as AI when it first appeared. Using a collections risk score (not to be confused with credit risk), companies have been able to rely on AI to adjust strategies and prioritize accounts for optimum results.
Let’s take a step back and review the evolution of collections over time. It started with companies using invoice value and age as the determining factor of how to prioritize accounts. This created a very cyclical return on results. One month, results look great because a large invoice was collected at the end of the month. However, the next month(s) didn’t look so great because the large invoice was collected by neglecting countless smaller invoices. These invoices add up over time and become increasingly difficult to collect as they age.
The next step in evolution was the introduction of strategies, which helped standardize the collections approach, spreading the focus equally among invoices. While results significantly improved with the introduction of strategies, they tended to plateau. Teams were left searching for how to capture the incremental improvements that would continue to improve cash flow. This is where AI was introduced. Looking at internal and external data sources, such as payment history and trade credit bureaus, an AI engine is able to predict the likelihood of a customer becoming delinquent 60 days in the future. Using that predictive view of accounts, the AI engine is able to assign a more granular risk profile and automatically adjust the strategies used for each customer and the prioritization of those accounts to prevent them from becoming delinquent. Introducing the AI engine provides companies with the incremental improvements that have been eluding them with previous processes.
AI in the dispute and deduction processes helps to identify and assign reason codes or categories allowing for root cause reporting and allows for prevention measures to be implemented. Additionally, the AI engine can automatically approve the disputes and deductions, based on predefined criteria, or automatically route them for resolution through advanced workflows. This accelerates the resolution cycle time and increases overall cash flow improvement.
Generally, deductions are identified during the cash application process. Leveraging AI during cash application increases the first pass hit rate of auto-applying payments to invoices. The AI engine is able to recognize remittance layouts to read and digitize the instructions for applying payments. Additionally, the AI engine improves hit rates over time as it learns from the exception processing of cash appliers. By monitoring how a user resolves an exception, the AI engine learns where the information used was located on the remittance and thereby learns how to apply future payments from that customer. By reducing the backlog of unapplied payments, all upstream processes (credit lines relieved timely, collection queues updated in real time, disputes resolved) benefit by creating more time for resources to focus on activities that drive cash flow.
Implementing AI in any area will help improve results. However, implementing AI across credit-to-cash will improve collections, credit risk, dispute resolution and cash application in addition to increasing increase cash flow. It will help your team uncover the incremental improvements that are sustainable over the long term.