In debt collection, the difference between a payment and a missed promise often comes down to one thing: the human connection between collector and consumer. What if you could predict which collector would build the best rapport with each account, before the call even begins?
That is what BestPair does. Our predictive pairing model learns each collector's strengths from the outcomes of their own calls, then routes each account to the collector whose profile gives it the best chance of paying.
The Problem with Next Available Collector or Queue Order
Most collection agencies still let the dialer hand each account to the next available collector, or work it in queue order. That approach treats every collector and every account as interchangeable, and behavioral science tells us otherwise.
Consider:
- Communication styles vary. Some consumers respond to empathetic, conversational collectors; others need a direct, no-nonsense one.
- Collector strengths are not uniform. One collector may excel with high-balance accounts that need negotiation, while another does best with first-time delinquencies that need a clear explanation of the options.
- Situation matters. The same consumer may respond differently to different approaches depending on where they are.
Next available collector or queue order ignores all of this. It is technically functional, but it misses the point.
How BestPair Predictive Pairing Works
BestPair's model works on three things:
1. Collector Profiling
We analyze each collector's historical performance to build a profile: rapport-building style, negotiation patterns and results across different account types. The profile is built from call outcome data and collector performance metrics.
2. Account Classification
Accounts are classified from the account attributes in the data you already hold. These signals help predict which collector profile will do best with each account.
3. Pairing
BestPair matches each account to the collector whose profile gives it the best chance of paying. It works with the dialer or collection system you already run, through a daily assignment file or agent queues.
How We Measure Results
Any vendor can show a before-and-after chart. We measure against a control group instead: one set of accounts is routed the way your floor routes them today, another is routed by BestPair, and the difference between the two is the result. It is clear, transparent and auditable, and it is how we test our own work.
Why This Matters for US Collection Agencies
Most agencies operate on thin margins. Collecting more from the same accounts, with the same collectors, changes the economics of the business: it is added revenue with no added headcount.
Your Controls Stay Yours
A common concern with AI in collections is compliance. BestPair was designed so that it changes as little as possible:
- Routing only. BestPair decides which collector takes which account. It does not contact consumers. It does not write scripts or change your call rules. It leaves the controls you run today for the FDCPA, Regulation F and the TCPA as they are.
- Auditable. Results are measured against a control group, so both sides can see exactly what changed.
Getting Started
Every engagement starts with a Floor Assessment: 4 weeks of analysis on your own call and payment data. It shows how much your results vary by collector and by account type, which is the spread predictive pairing works on. If the data shows no spread between your collectors, the fee is refunded in full.
If you want to know what your own data says, book a call with Kevin.