You’ve heard AI can improve collections workflow. But not all AI machine learning technologies for the accounts receivable management (ARM) industry are alike. There’s AI in accounts receivable, and then there’s dynamic scoring using “explainable” AI (xAI)—a true industry breakthrough.

Unlike static account scoring that updates only periodically, dynamic scoring draws upon a wide array of alternative data sources to create comprehensive consumer profiles. xAI makes machine learning and artificial intelligence transparent and understandable to the human eye, eliminating the blackbox model of traditional AI decision making—along with regulatory concerns over the use of AI in accounts receivable.

For third-party collection agencies, dynamic scoring plus xAI is a powerful combination that can drive unprecedented collections results without complicating compliance management.

Recently, I sat down with Ontario Systems Principal Data Scientist Ian Baldwin to discuss why dynamic scoring using xAI is unique, how it works, and what it means for the ARM industry. This blog post offers some highlights of our discussion. To hear the full webinar, “Collect Smarter: How ‘Explainable’ AI Is Transforming the ARM Industry,” you can access the free on-demand recording here.

How Dynamic Scoring Revolutionizes Consumer Segmentation for Third-Party Collection Agencies

Rather than generating a simple one-time snapshot of consumers, dynamic scoring involves an ongoing analysis of information from both standard sources and thousands of untapped alternative data sources.

Driven by machine learning, this process delivers big-picture insights by probing consumers’ digital presence, spending habits, psychological behavior, engagement history, educational levels, employment history, mortgage records, driving records, home ownership, and more. It goes even further, taking into consideration external influences that can impact consumers (the COVID-19 pandemic, for example) as well as individual and collective agent behaviors.

The result is better prioritization of consumer accounts. Aside from the obvious accounts with large dollar amounts and highest scores on paper, this enhanced segmentation allows businesses to recognize existing opportunities in otherwise stagnant aspects of their portfolios.

 

“Traditional, static scoring models do not update fast enough to reflect the swirling mass of uncertainty today’s consumers live in.” – Ian Baldwin, principal data scientist, Ontario Systems

Optimized Segmentation Offers Meaningful, Actionable Intelligence

With more data input, dynamic scoring goes beyond traditional segmentation. Algorithms in the machine learning models increasingly develop, gain experience, and become more effective in communicating with one another.

Set up correctly, the algorithms assist the machine learning models while working together to predict a variety of different solutions that are beyond human intelligence. The resulting propensity-to-pay scoring and creation of dynamic consumer personas inform dialer logic.

In this ongoing process, machine learning continuously analyzes new data, and AI recommends next steps. In lieu of incremental batch updating or costly in-depth analysis, accounts are automatically refreshed. For the first time, ARM businesses can be truly nimble, using current, broad-based data to derive meaningful insights into their scoring model, methods, employees, and types of accounts.

xAI: Artificial Intelligence That’s Built for Compliance and “Human Aware”

Without transparency, blackbox technology is extraordinarily dangerous from a compliance standpoint because it hides its critical calculations. xAI resolves the problems of AI combined with blackbox machine learning by effectively addressing consumer debt for all stakeholders in a completely transparent way. Put another way, xAI’s decision making is “human aware.”

Rather than having the tools make its own decisions, xAI serves as an aid to human decision makers. Pulling back the curtain is SHapley Additive exPlanations (SHAP), which explains individual predictions. Users can identify which specific factors (current balance, digital presence of the consumer, vehicle ownership, and mortgage, for example) have changed the scoring decision and exactly how much.

With SHAP, xAI reveals an explainable picture of the model to collection agencies and to regulators. It also provides information needed to refine the models, do internal regulatory checks, and protect against discriminatory practices.

An Easy Way to Unlock Your Company’s Potential for Growth

Clearly, the ARM industry has much to gain by embracing machine learning and artificial intelligence—particularly dynamic scoring coupled with xAI. It’s better for the consumer, better for business, and a better way to operate in a highly regulated environment.

Agencies that leverage dynamic scoring with xAI are better equipped to accelerate and grow because they better understand their own business as well as consumer behaviors, communication preferences and, most importantly, the best times to deploy those communication attempts. These agencies can also recover more revenue with fewer manual touches.

As dynamic scoring with xAI develops further, it can create a system fully embedded in the workflow, helping make day-to-day operational decisions and responding to events in real time. As the models learn over time, the workflow can change to trigger the appropriate action without collectors having to monitor or change systems routinely. Early adoption of dynamic scoring using xAI will no doubt give third-party collection agencies a distinct market advantage.

Learn More About Dynamic Scoring Using “Explainable” AI and Its Future in Accounts Receivable

Want to better understand the benefits of dynamic scoring and xAI and how they compare with other AI technologies for the ARM industry? Listen to the full on-demand recording of “Collect Smarter: How ‘Explainable’ AI Is Transforming the ARM Industry.” Access your free copy today.