Stop the hiring craze: How TrueAccord reached 30,000 cases handled per agent

By on December 1st, 2015 in Industry Insights, Machine Learning

Hiring collectors is one of the biggest challenges in collections and recoveries. Most collection processes are manual and require extensive training: the low, commission based salaries and the adversarial nature of the job lends itself to high turnover. Once agents are trained and working, collection managers must keep them up-to-date on debt collection regulations, as well as implement quality controls. One ill-trained agent’s compliance violations can put an operation at risk due to the many regulations that govern the US collection and recovery process. This results in a big investment in human resources. It’s no surprise this is one of the biggest line items for any collection operation. Beyond payroll and training, turnover results in low morale, often hurting organizational cohesion.

TrueAccord agents aren’t traditional collectors. While trained in every aspect of collection compliance, this team of experts is tasked with resolving complex cases and tuning our machine learning system. Their compensation doesn’t have a commission component. The team’s name, Customer Engagement, hints at its consumer-first focus. Even with a highly qualified group of agents, growing linearly with placed volume can still hurt the company’s cost structure and prevent TrueAccord from continuing to provide long-tail or low-balance collections. We realized we needed to find a scalable way to grow, so approached the solution in two different ways.

Scaling with decision automation

The technology platform that underlies TrueAccord’s system is programmed to replace many human decisions with machine-based ones. Replacing the call center paradigm with a machine based one has many benefits, chief among them is, far less need for humans to be involved in the process. The patent-pending engine that powers the collection platform, Heartbeat, tracks consumer interactions with collection communications; notes their browsing pattern on the website and uses historical data to decide what its next action should be. Heartbeat handles follow-ups on failed payments and promises to pay, triggering communications on its own. It even asks for an outbound call, when deemed to be the best course of action.

In this model, agents stop being collection tacticians making micro decisions while handling talk-offs. They are, instead, collection strategists – deciding what the general course of action should be, resolving difficult cases and teaching the machine how to improve. This is a fundamentally different way to handle collections – from human-heavy to expert-based automation.

Scaling with process automation

Another advantage of a machine-based, digitally focused system is streamlined process automation. Dispute handling is a great example of this in practice. Traditionally, disputes under FDCPA guidelines are received in writing, via postal mail, within a certain window of time after the initial consumer communication. This highly manual and inefficient process; consumers mail disputes that may get lost in the mail or through paper-heavy process and turnaround is, at the very least, several weeks. Given how expensive it is to process disputes, collectors sometimes choose to discourage them, as much as possible, while processing the minimum required by law. The danger of this strategy is that the savings in manpower may be eclipsed by the cost of legal violations.

Compare that process with eDisputes. Since the platform is automated, adding a conspicuous link to an online dispute experience is very easy. Consumers are asked a few questions, identify their issue and get an immediate notification of a dispute receipt. If required, the client is pinged for verification, after which they can upload documents via the platform and those are then presented to the consumer in an easy to understand interface. In-house agent involvement in this process has been reduced to almost zero. And indeed – the percent of disputes filed via direct mail has dropped by almost 100%, while consumers get a more compliant, easy to access response to their dispute.

Bottom line – getting to 30k

TrueAccord considers technology first. We set out to build a platform and system that continues to scale as more clients trust us with their business. We invested in a platform that relies on expert agents rather than a large call center, using data to replicate their best decisions at scale. Then, based on our evolving understanding of collection needs and how consumers use the platform, we continue to add automation to replace work that was previously manually done. In 2014, each of our Customer Engagement agents handled 10-15k active cases; this year, the number has already surpassed 30k, yet we are far from perfect optimization. This means we’ll continue to be able to offer clients flexible collection plans, handle low amount accounts and high-touch accounts effectively. We are poised to manage attractive pricing models. And, as a bonus, a system that involves less human-to-human interaction is leaps and bounds more compliant than one that relies on commission-based phone calls. It’s a win for all parties involved.

Leave a Reply

Your email address will not be published. Required fields are marked *