More collections, fewer pushbacks — ARM case study

How a leading U.S.-based ARM firm doubled agent productivity, added more than $1M in monthly collected revenue, and cut Do-Not-Call requests by 37% — in six months with Sedric’s Real-Time Agent Assist.

A leading U.S.-based accounts receivable management (ARM) firm deployed Sedric’s Real-Time Agent Assist across its collections floor in May 2024. Six months in, the program had materially shifted both the unit economics and the compliance posture of the operation.

Agents on the assisted floor are now completing roughly twice as many calls per month, with average call duration down nearly 29%. Payment-related call dispositions — full and partial payments, promises to pay, and structured payment arrangements — are all up sharply, with a combined monthly revenue impact in excess of $1M on a measured basis and several million more in modeled upside.

Crucially, those collection gains were not bought with more aggressive calling. Over the same window, the share of calls ending in a Do-Not-Call request fell by 37% — a leading indicator of complaint risk under Regulation F and the FDCPA. In other words, agents had more, better, more compliant conversations at the same time.

The business challenge

ARM operators run a margin-thin, conduct-regulated business at high call volumes. Three structural pressures defined the brief that operations and compliance leadership took into the project.

1. Coverage and consistency on the call floor

Quality assurance in collections is typically sample-based, retrospective, and uneven across teams and shifts. Once an agent has said something problematic — or, just as often, missed a compliant opening to advance the call — the call is over, the disposition is logged, and the consumer’s experience is set. Post-call QA can catch patterns, but it cannot recover lost revenue or undo a complaint trigger.

2. Regulation F operationalization

The CFPB’s Regulation F has translated long-standing FDCPA principles into very specific operational expectations on call frequency, communication channels, validation, and consumer opt-out. Every call now carries dual obligations: convert the contact into payment when possible, and stay inside a tightening conduct perimeter. Agents have to navigate both in real time, on every call.

3. Productivity ceiling and complaint risk move together

The instinctive lever for collections lift — more calls, more pressure — is also the lever that drives complaint volume, Do-Not-Call requests, and CFPB exam exposure. The firm needed a way to lift productivity and conversion without trading them for complaint risk and reputational damage.

Why Sedric Real-Time Agent Assist

Sedric’s Real-Time Agent Assist runs alongside the agent during the live call, surfacing in-the-moment guidance grounded in the firm’s own policies and the regulatory expectations applicable to consumer collections. Rather than scoring calls after the fact, it nudges the agent toward the next best action while the conversation is still in motion.

For this deployment, the firm prioritized four capabilities:

• Live policy coverage. Continuous monitoring against the firm’s collections policies and applicable regulatory standards, with prompts surfaced to the agent in real time.
• Next-best-action guidance. Conversational prompts that help agents handle objections, position payment options, and close on a promise-to-pay or payment arrangement without crossing conduct lines.
• Compliance guardrails. Pre-emptive alerts on script deviations, prohibited phrasing, and call-handling patterns that typically precede a Do-Not-Call request or written complaint.
• Disposition-level analytics. Visibility for team leads and the second line into conversion patterns by disposition, agent, and shift — to direct coaching where it moves the most revenue.

Implementation

Sedric was deployed on the firm’s live collections floor in May 2024. Within the first measurement window, the platform was integrated with existing telephony and dialer infrastructure, calibrated to the firm’s collections policies, and embedded into the agent desktop with targeted enablement for team leads and the front line. No additional headcount was added to the compliance or operations function during the measurement period.

Outcomes (June → December 2024)

Six months in, the program had moved every line in the call-floor P&L in the same direction at the same time: more output, better unit economics, and a measurably lower complaint-risk profile.

Collections lift

Payment-related call dispositions are up across the board against the June baseline, translating into more than $1M in measured monthly revenue impact and several million more in modeled upside as the lift compounds.

• Full or partial payment (segment A): 0.68% → 0.86% · +26% · +$156,684 monthly revenue impact*
• Full or partial payment (segment B): 1.33% → 2.32% · +74% · +$872,265 monthly revenue impact*
• Promise to pay (segment A): 6.09% → 9.03% · +48% · +$2,600,345 modeled monthly upside*
• Promise to pay (segment B): 0.04% → 0.06% · +41% · +$14,574 modeled monthly upside*
• Payment arrangement: 0.05% → 0.77% · +1,354% · +$637,424 modeled monthly upside*

Agent productivity (August → December 2024)

• Calls per agent, per month: +225%.
• Average call duration: −29% — agents are handling shorter, sharper conversations and reaching more consumers per shift.

Compliance posture and complaint risk

• Do-Not-Call requests as a share of total calls: 0.46% → 0.29% — a 37% reduction over the same window in which collections lifted.
• DNCs and consumer pushbacks are widely treated as leading indicators of complaint volume and CFPB exam exposure. The directional move on this metric, in combination with the productivity and revenue lift, is the result that distinguishes this deployment from a pure productivity play.

What the data is actually saying

The standard concern about putting AI on a collections floor is that productivity gains will come at the cost of conduct — more aggressive calling, more cut corners on disclosures, more complaint volume. The June-to-December data shows the opposite. The same intervention that doubled call throughput and lifted conversion across every payment-related disposition also reduced the rate at which consumers asked not to be contacted again. The most plausible explanation is the simplest one: in-call guidance helped agents have better conversations — more compliant and more effective at the same time.

*Revenue impact methodology: monthly revenue impact and modeled upside are calculated as the change in the share of calls with a given payment-related disposition, applied to the firm’s monthly call volume, valued at an assumed average debt balance of $500 per converting call. Figures rounded to nearest dollar. Productivity metrics measured August through December 2024. The firm’s identity is withheld at the customer’s request.