For most of the last few decades, collections agencies and in-house collections professionals have been graded on activity and behavior.
Calls placed. Accounts touched. Contacts attempted. Hours logged. If it moved, it counted.
Technology changed. The scorecard didn’t.
AI didn’t fix that problem. It put it on steroids.
Today, agencies proudly report how many AI calls were made, how many conversations were “handled,” how fast responses were, or how much volume the system pushed through. It all looks very modern. Dashboards light up. Charts go up and to the right.
And almost none of it answers the only question creditor clients or CFOs actually care about:
Did it recover cash in a predictable, compliant, and cost-efficient manner all while delivering a fair, respectful consumer experience?
Creditors don’t buy activity. They buy outcomes.
They care about recovery dollars, promises kept, cost-to-collect, complaint exposure, and consistency across portfolios.
The ACA has reported that the ARM industry returns tens of billions of dollars annually to creditors, over $90 billion in prior studies. That’s the benchmark. Cash recovered. Not conversations initiated.
Here’s the core truth most agencies still resist: activity is a cost. Outcomes are the product. Neglect that shift, and no amount of “advanced AI” talk will save client relationships.
The activity metrics illusion
According to a CFPB study, 79% of consumers who filed a debt collection complaint said they had first tried to resolve the issue directly with the company. But by the time a complaint hits a regulator, trust has already broken down.
Creditors know this, and they track which agencies accelerate that breakdown.
The problem for most collection agencies and internal collections teams? It’s that the model they follow once worked perfectly.
That model was created in a world where people were the bottleneck. Everything, from staffing models to reporting frameworks, evolved to manage their work efficiently because there was usually a lot of it to do.
Calls per hour. Contacts per agent. Handle time. Utilization. Those metrics still help run a floor. They do almost nothing to explain performance to a creditor.
AI didn’t break that habit. It reinforced it.
Suddenly, agencies could show bigger numbers faster. Volume became easier to sell than effectiveness. And internally, it felt good. Progress without friction.
But creditors aren’t impressed by volume anymore. They assume it. What they increasingly demand is control and accuracy.
What illustrates this? The ACA has noted the industry’s accelerating shift toward digital engagement and self-service models. This modernization isn’t about creditors wanting more attempts. They want more control and more resolutions.
Regulation is making every interaction crucial
Another wrinkle? Regulators are putting the brakes on excessive activity.
Under Regulation F, calling about a particular debt more than seven times in seven days (or calling again within seven days after a conversation) could constitute harassment.
Plus, the FCC has put limits on prerecorded calls.
So volume-first strategies are running straight into a legal ceiling. This points toward the need for smarter targeting and better outcomes.
Call-frequency limits aren’t theoretical guardrails. They’re enforceable thresholds. In a capped environment, quality of interaction replaces quantity as your growth lever.
What creditors actually care about (still!)
Creditors have always cared about outcomes.Banks, fintechs, utilities, healthcare providers, and telcos are all operating under tighter regulatory scrutiny, thinner margins, and higher reputational risk.Enterprises don’t want collection vendors or teams who “try hard.” They want partners who can explain results and execute accurately.The outcome metrics that actually matter haven’t changed much. So you should build an AI scorecard around them to measure actual success:
- Net recovery rate by segment. Early, mid, late, purchased debt; blended totals hide reality.
- Promises kept, not just promises made. PTP rate without kept performance is not useful.
- Cost per dollar recovered. Automation is only progress if margins improve.
- Resolution velocity. How quickly accounts move from contact to payment.
- Complaint rate per 1,000 accounts. Volume without control shows up here first.
- Predictability. Creditors defend stable curves, not sporadic spikes.
These are the numbers creditors defend internally. These are the numbers regulators look at. These are the numbers that decide renewals and and define pricing.If an agency can’t clearly connect its activity to these outcomes, it’s living on borrowed time—no matter how sleek the tech stack looks.
Why AI makes outcome measurement mandatory
Consider this too-common scenario:
- An agency triples outbound AI call volume across a portfolio.
- Right-party contact ticks up a bit. Promises-to-pay remain flat. Kept-promise rate declines because targeting logic didn’t change.
- Complaints increase due to higher contact frequency.
- Gross collections look busy, but net recovery and cost-to-collect are unchanged or worse.
Activity surged. Economics didn’t. That’s the danger of measuring activity instead of resolution.AI should be evaluated like any other investment: by ROI.
Making the shift without breaking client relationships
Some collections agencies avoid outcome-based reporting because they’re scared of client pushback.
Frankly, they’ve got it backwards. Creditors already think this way. Agencies are the ones clinging to old scorecards.
The shift doesn’t require ripping out legacy metrics overnight. It requires reframing them. Activity should support the outcome story, not replace it.
The smartest collections teams start by pairing traditional reports with outcome-driven analysis showing how strategy, timing, and channel mix connect to recoveries. Over time, the conversation naturally changes.
Smart collections teams don't just report on the past; they connect recoveries directly to the "why," factoring in strategy, timing, and how they reached out.
This moves the whole conversation in the right direction.
Clients stop asking how many calls were made. They start asking why results improved…or didn’t.That’s not a threat to an agency’s business model. That’s proper alignment with client need.
The industry and regulators are sending clear signals
This shift isn’t theoretical. It’s already happening.
Public complaint data, particularly through federal and state channels, continues to show that high activity does not equal low risk.
For instance, the CFPB received over 207,800 debt collection complaints in 2024. That’s almost double the year before.
One possible reason? Analysis of that data shows 45% of those complaints were related to debt the consumer said they did not owe.
In other words, activity doesn’t correlate with control, but can actually amplify errors and complaints.
At the same time, delinquency pressure is rising across multiple asset classes. ACA reporting points to student loan delinquencies above 10% in recent periods. More volume is coming, but tolerance for inconsistent recovery performance is shrinking.
Credit-washing fraud is spiking, too. So “more calls” can mean “more exposure” unless your system is outcome-driven and controlled.
It’s why collections must move toward measurable, defensible, outcome-based performance.Agencies that adapt early gain leverage.
Agencies that don’t usually adapt later—under pressure, on someone else’s timeline.
A quick word on Overtime’s approach
Overtime was built around this reality from day one.
Rather than optimizing for volume, Overtime’s voice-first AI agents are designed to pursue specific recovery outcomes, like securing a promise-to-pay or resolving an account, within strict compliance and brand guardrails.
Every interaction runs on consistent logic and is fully auditable, which means agencies can measure success by results achieved, not just conversations initiated.
That makes recoveries more predictable, reporting more credible, and client conversations grounded in metrics that actually matter.
To sum up? Results are the product
AI didn’t make collections more complicated. It made the old measurement model obsolete.
Agencies that continue to focus on volume will struggle to differentiate, retain clients, or defend their value. Agencies that anchor everything to outcomes gain clarity, credibility, and control.
In modern collections, activity is just the input. Results are the product.
The collections professionals who understand this will define the next decade of recovery.