Artificial intelligence (AI) is generating a lot of interest in the accounts receivable management (ARM) industry. It’s also raising concerns about potential liability. ARM leaders recognize the benefits AI can bring, but they’re hesitant to take that big leap into the unknown.
Yes, the typical “black box” AI approach is unknown, and unknowable, to most users. Fortunately, explainable AI—a completely transparent model that’s tailor-made for the highly regulated ARM industry—offers all of the collecting power and efficiency agencies want and none of the perceived risk that’s holding them back.
Recently, I sat down with Ontario Systems Senior Director of Data Science and Product Management Greg Allen to discuss why explainable AI is the best way forward for debt collection agencies concerned about Regulation F, rising costs, and untapped revenue.
Why Explainable AI Technology Is a Must-Have for Third-Party Collectors
Let’s start with a brief overview of how the technology works. As machine learning combs consumer data (both proprietary and alternative) to detect patterns with increasing accuracy and precision, explainable AI develops deep consumer profiles. A dynamic scoring model updates propensity-to-pay scores continuously, rather than every 30–60 days as typical static scoring models do.
With an up-to-date, 360-degree view of the consumer, collectors can optimize portfolio segmentation and contact strategies on an ongoing basis. As a result, they can extract more and more value, even among lower-priority accounts, with fewer contacts.
Here’s what this collections “superpower” will enable you to do.
1. Make the best use of limited contacts
The CFPB new rules for debt collection, along with some state laws, limit call frequency and impose certain restrictions on the use of text and email. With so little room to maneuver, ARM agencies can no longer afford contacts that don’t achieve a desired result such as authorization for digital communications or setting up a payment plan. Explainable AI can help you make every contact count, giving you the insights you need to communicate more prudently across the board.
2. Maximize employees’ time
Explainable AI technology transforms the segmentation process by offering a far better understanding of consumer behavior, preferences, and financial circumstances. Accounts are placed in the right buckets so the proper contact strategies are applied to all types of accounts, while AI-driven automated workflows ease daily burdens. Both benefits empower agents to focus solely on what will produce results.
3. Create less friction for consumers
Ensuring a positive experience for consumers helps to improve client satisfaction, minimize litigation risk, and optimize revenue recovery. With explainable AI, you’ll understand consumers’ preferences for specific communication channel(s) and times of day you’re likeliest to get a favorable response. Respecting these preferences reduces complaints and makes consumers more inclined to pay quickly or work out payment arrangements.
4. Become an intelligent, data-driven enterprise for long-term success
A data-driven foundation supported by machine learning and AI will allow your business to adapt readily to changing regulations, consumer expectations, and market pressures. Rather than reacting to new problems as they emerge, you’ll have the visibility you need to avoid disruption, adjust course, and continue advancing your business goals.
Why Explainable AI is Safe to Use for Debt Collection
Because it’s completely transparent and easy to interpret, explainable AI eliminates the fears and risks that a “black box” model can invite. If the thought of introducing explainable AI technology makes you feel queasy, here are three reasons why it shouldn’t.
1. Regulators approve of AI and are planning for widespread industry adoption
One of the biggest prevailing myths is that ARM industry regulators are opposed to using AI for collections. Far from it: the CFPB recently solicited comments related to AI use, with proposed rules to follow. The Federal Trade Commission has released its own set of guidelines around truth, fairness, and equity in the use of AI.
In the interest of preventing biased practices that harm consumers (regardless of intent), regulators are most concerned with how data is collected, how it’s used, and how it’s controlled and managed. Explainable AI will allow you to understand and demonstrate all of this in detail.
2. Data privacy and risks needn’t be a concern
Regulators are also concerned about data security and protecting consumer privacy. But there’s nothing inherent in the use of AI that increases a company’s exposure to liability or risk. You just need to make sure you have all the proper protocols in place.
3. Explainable AI can be audited to prevent biased collections activity
AI can be biased, just like humans. Fortunately, explainable AI offers “look back” testing, allowing you to uncover hidden biases in the algorithms driving your workflows. With periodic testing, you can ensure you’re using the right kind of data and your agents’ daily activities aren’t resulting in disparate impacts on protected classes of consumers.
Rozanne Andersen is the Chief Compliance Officer at Finvi. Greg Allen is the Senior Director of Data Science at Finvi.
This article was originally published on Finvi's corporate blog and is reprinted here with permission.