Compliant B2B invoice follow-up: how much do you actually keep at an 18% success fee?

The cash-flow math, without the sales gloss

Most credit and AR teams I talk to don't need a philosophy lecture. They need to know what lands in the bank. So take a $25,000 aging invoice that's 97 days past due.

At a 45% traditional agency fee, a full recovery nets you $13,750. At 18%, you net $20,500. That's a $6,750 difference on one invoice. Multiply that across a portfolio of, say, 40 stuck invoices averaging $6,000 each, and a 60% recovery rate looks like this:

• Total placed: $240,000
• Recovered at 60%: $144,000
• Net at 45% fee: $79,200
• Net at 18% fee: $118,080

That's a $38,880 swing — roughly the cost of a full-time AR analyst — from fee structure alone. No new customers, no new sales, just a different split on money you already earned.

The catch worth naming: a lower fee usually means a lighter-touch workflow. An agency at 45% is sending demand letters and making phone calls with a collections posture. A payment resolution service at ~18% is doing structured follow-up, payment plans, and dispute handling. Whether the lower fee nets you more depends on your portfolio's recovery rate, which you should test on a sample before you move everything.

What "compliant" changes about your follow-up

Compliant invoice follow-up much of the time comes down to three practical constraints: who you're contacting, what you say, and what you can prove you said.

For B2B commercial debt, you generally have more latitude than with consumer debt, but "commercial" is narrower than people assume. A sole proprietor disputing a $4,000 invoice can push you toward consumer-style rules depending on your jurisdiction and the nature of the debt. A finance director at a 200-person company is a different conversation. Your follow-up cadence, tone, and documentation should reflect that difference, not treat every past-due account the same.

Three things a compliant workflow does that a sloppy one skips:

1. Accurate records. Every contact attempt, every promise to pay, every partial payment, logged with a timestamp. If a dispute escalates, this is your defense.
2. Clear disclosure. If you've placed the account with a third party, the debtor should know who they're dealing with and how to dispute the debt.
3. No misrepresentation. Don't imply legal action you won't take. Don't threaten credit reporting you can't actually file. That's where agencies get sued, and where creditors get dragged in.

None of this is legal advice — your counsel should sign off on templates. But operationally, a compliant follow-up sequence is mostly about discipline and documentation, not legal firepower.

Where AI helps and where it doesn't

AI is good at the boring middle: sequencing follow-ups, drafting tone-appropriate emails based on the debtor type, flagging accounts where the payment pattern changed, and keeping the record clean. It's not good at the last mile. When a controller says "we can do $3,000 now and $2,000 a month," a human needs to make that call, confirm it in writing, and decide whether the plan is worth accepting versus escalating.

A workable pattern looks like: AI handles the first four touches and the documentation. A human handles anything that turns into a negotiation, a dispute, or a payment plan. That division keeps your success rate up without pretending software can close a difficult account on its own.

If you want to see how a no-recovery-no-fee structure works in practice, the unpaid invoice recovery setup here is a reasonable reference point for how the 18% split is applied.

Underwriting the account before you chase it

Here's a piece most fee comparisons skip: not every past-due invoice is worth the same effort. A $2,000 invoice from a company that's clearly still operating and just slow-paying is a follow-up problem. A $2,000 invoice from a company showing distress signals is a recovery problem — and sometimes a write-off problem.

Traditional bureau scores don't tell you much about a private B2B debtor's current ability to pay. They lag. A risk score built from payment behavior, aging patterns, and recent activity on the account itself tends to be more useful for deciding where to spend human effort. If you're running a portfolio, segmenting by that kind of signal before you place accounts is where the real margin sits — not in shaving another percentage point off the fee.

If you want to look at how that scoring is structured, there's a summary at the risk intelligence page.

A quick word on tooling and integration

If your AR data already lives in an ERP or a billing system, you don't want a recovery partner that requires manual CSV exports every month. The practical question is whether the workflow can read your aging report and write back payment status without someone rekeying it. That's usually an API or MCP question, and it's worth asking before you sign anything. Docs for that kind of integration typically live alongside the product — see the creditor API documentation for one example.

The reason this matters for the fee math: every hour your team spends reconciling spreadsheets is a real cost that doesn't show up in the 18% versus 45% comparison. A slightly higher fee with clean integration can beat a lower fee with manual overhead. Run the numbers with your own labor cost, not just the split.

So what do you actually keep?

At an 18% success fee with no recovery, no fee, you keep about 80% of whatever is collected. On a $10,000 recovery, that's roughly $8,200. On a $144,000 portfolio recovery, that's roughly $118,000. The exact number depends on your recovery rate, your mix of accounts, and how much of the work your team still has to do.

The honest framing: a lower fee is only better if the recovery rate holds up. Test it. Place a sample. Compare net dollars, not percentages. That's the only comparison that survives contact with a board meeting.