Resolve unpaid invoices: net cash at an 18% success fee vs traditional collection agency rates

Start with net recovery, not gross

Credit and ops people talk about recovery rates. Finance talks about net cash. Those aren't the same number, and the difference is the fee.

Take a $10,000 B2B invoice sitting at 95 days. Two outcomes:

• Traditional agency at 42%: recover $10,000, pay $4,200, net $5,800.
• Success fee at 18%: recover $10,000, pay $1,800, net $8,200.

Same effort on the debtor, same cash in. A $2,400 difference per $10k — about 41% more net cash on that file. Spread that across a quarter of slow-paying accounts and it stops being a rounding error. That's the main reason a collection agency alternative with a lower fee is worth a look before you sign a placement agreement.

Two caveats, because I'd rather you trust the number than the pitch:

• Fees only matter on dollars actually collected. A 45% agency that recovers where an 18% service doesn't is still the better trade on that specific file. Compare net, not rate.
• Partial recoveries and payment plans change the timing. A $10k invoice settled for $7k over four months nets differently than a lump sum. Run it both ways.

Why the fee spread exists

Traditional agencies carry a cost structure built on phone floors, skip tracing, and legal escalation. That's expensive, and it's priced into 40–45%. It also means they tend to take accounts in bulk and work them in queues.

A payment resolution service built on AI plus human workflows pushes the low-complexity work — reminders, payment links, plan setup, dispute intake, documentation — into software, and routes only the accounts that need a person to a person. Lower cost to serve, lower success fee. That's the honest version of why 18% is possible. It isn't magic; it's different unit economics.

It also means the model works better on some files than others. Straightforward past-due commercial invoices with a reachable AP contact? Software handles most of it. Contested invoices, insolvency signals, or a debtor who's gone quiet for six months? Human judgment earns its keep, and no fee structure fixes a bad debt.

A quick cash-flow frame for deciding per invoice

I wouldn't set one policy for the whole ledger. Segment by age, amount, and dispute status, then ask three questions:

1. What's my internal cost to keep chasing? At some point your AR person's hours cost more than the expected recovery. That's the point to hand off.
2. What's the realistic recovery probability? A fresh 35-day invoice and a 200-day invoice are not the same asset. Price the effort accordingly.
3. What's the net at each fee? If the expected recovery is $6,000, compare $6,000 minus 42% against $6,000 minus 18%. Decide on that.

For most teams, the sweet spot is: internal follow-up through day 60–90, then route the stubborn files to a success-fee service. You keep the easy money in-house and stop funding a collections desk for accounts that need a different approach. If you want the longer version of how that handoff works, unpaid invoice recovery that keeps you ~80% walks through it.

Where AI + human actually helps

Software is good at consistency. It doesn't forget to follow up on day 61, it logs every contact, and it can run a payment plan offer across 400 accounts the same afternoon. Humans are good at the edge cases: a debtor disputing line items, a partial settlement that needs sign-off, a relationship you don't want to burn.

In practice, a resolution workflow looks like this:

• Automated reminders and payment links on a cadence, with clear opt-out and contact rules.
• Structured intake for disputes, with documentation collected up front.
• Payment plans and settlements negotiated inside limits you set.
• Escalation to a human when the account warrants it — or when it's time to write it off and move on.

Compliance isn't a feature you bolt on later. Commercial and consumer recovery both sit under real rules, and the practices differ by jurisdiction and by whether the debtor is a business or a person. A "no recovery, no fee" structure doesn't excuse sloppy contact practices. Ask any provider how they handle disputes, validation, and recordkeeping before you place a file.

Screen before you extend, recover after

Recovery is the expensive way to learn a customer was a bad risk. Underwriting is the cheap way. If you're extending terms to new B2B accounts, the credit decision up front matters more than any collection rate later.

Traditional bureau scores were built for consumer lending and don't always translate to a small commercial buyer with thin files. A risk score built on payment behavior across receivables — the kind behind the FP Risk Score — can flag accounts before they age. Pair that with terms that match the risk: deposit, shorter net terms, or a card on file. Then recovery becomes the exception, not the operating model.

If you're wiring this into an existing stack, most of it is API-shaped. Creditors running their own systems can pull risk and resolution actions through the MCP and API tooling rather than logging into another portal. If you'd rather start with the software side, the main platform is at fixpayment.org.

The bottom line

An 18% success fee versus 40–45% isn't a marketing angle — it's a cash-flow decision you can model on one spreadsheet. On a $10k file, it's roughly $2,400 more net. On a quarter of slow payers, it changes what you can reinvest.

Be clear-eyed about the tradeoff: fee structures don't create recoveries. Probability does. Route the files that need human negotiation to humans, let software handle the rest, and measure net cash — not gross recovery, not the fee percentage in isolation. And screen new accounts so you're solving fewer of these problems next year.

Educational commentary on receivables and recovery operations, not legal advice. Practices must follow applicable consumer and commercial rules.