Insurance Payment Processing Companies

Two money flows that only look like one business

Almost every buying mistake in insurance payments starts with the assumption that a carrier or an agency has a payments problem. It has two, and they behave nothing alike. Money comes in from policyholders and producers on a schedule set years in advance, in small amounts, from people the insurer already knows. Money goes out to claimants on no schedule at all, in amounts that can be four orders of magnitude larger, to payees the insurer met three days ago.

A failed inbound payment is not an abandoned cart. It is a lapsed policy — an uninsured driver, a denied claim, a regulatory complaint, and a customer the insurer paid an acquisition cost to win and has now lost. A delayed outbound payment is a claimant who tells everyone. Neither is a payments metric, which is why buying insurance payments on price per transaction produces bad outcomes.

Premium collection is a billing problem wearing a payments costume

Premium billing looks like subscription billing and is not. A policy has a term, an installment plan that may carry its own filed fee, mid-term endorsements that change the amount owed, commercial-lines audits that true it up afterwards, and a cancellation process governed by statute rather than by the vendor's dunning settings.

Three collection models coexist, each pushing the payment problem onto a different party. Under direct bill the carrier bills and collects, owning the experience, the card cost and the lapse risk. Under agency bill the producer collects, keeps commission and remits net, so the money sits with the agency — where fiduciary accounting rules bite. Under list bill an employer remits for a roster in one payment, and the allocation across policies is the hard part.

Dunning has to follow the cancellation timetable, not the retry schedule

Off-the-shelf platforms ship a sequence tuned for e-commerce subscriptions: retry on day one, day three, day seven, then suspend. Insurance has a statutory grace period — a window after the due date during which coverage remains in force despite non-payment, its length set by state law and policy form and differing across life, health and property lines — and a notice-of-cancellation requirement with its own minimum days and proof-of-delivery rules. The two ladders have to be the same ladder, or the carrier either cancels earlier than the law allows or leaves an unpaid policy in force until a claim exposes it. Stored-card decay causes the same failure quietly, which is why account updater services and network tokenization — the network pushing refreshed credentials, and a token it keeps current in place of the card number — matter more here than elsewhere.

Bank debit changes the failure profile rather than removing it. Pulling premium directly — ACH debit in the US, Bacs in the UK, SEPA Direct Debit in the eurozone — eliminates card expiry, which is why it dominates European premium collection and why a mandate-based collector such as GoCardless suits it. It does not eliminate reversal risk: the UK Direct Debit Guarantee gives the payer an immediate refund on request with no time limit and no right of reply, and SEPA Core allows an unconditional refund for eight weeks. Cheaper than cards on high-ticket premium, and less final.

Who pays the card fee, and why the answer changes at the state line

Every biller-direct platform sold into insurance makes the same offer: absorb the cost of accepting the payment, or pass it to the policyholder as a separate line. The vocabulary matters, because the legal treatment follows the label.

Insurance then sits under a second regulator on top of the networks, which is where general-purpose payments advice fails. In several states a charge added to a premium payment is treated as a charge for insurance: it may have to be filed, it may collide with anti-rebating and unfair-trade-practice statutes, and it may be barred outright as a "pay-to-pay" fee. Paymentus, an NYSE-listed electronic bill presentment and payment company with a large insurance book, states in its own Form 10-K that several states have provided guidance or prohibitions against the use of convenience or similar pay-to-pay fees in certain industries, and names billers' ability to pass fees on as a risk factor. When the vendor puts the risk in its securities filing, a carrier should not treat it as settled.

So the fee-bearer decision is a compliance and filings decision that finance executes, not a finance decision compliance reviews. A national carrier usually ends up with a state-by-state matrix in which the same platform runs absorbed in some states and payer-funded in others, and the platform must support that natively — InvoiceCloud and Paymentus both do, and that flexibility is much of what a carrier buys from an EBPP vendor rather than from a plain acquirer.

Agent and broker premium trust accounting

When a producer collects premium, that money is generally not the agency's money. In most states it is fiduciary property held for the carrier and, in some circumstances, for the insured, and it must sit in a designated premium trust account that is not commingled with operating funds. Commingling is a licence-level offence, not a bookkeeping irregularity, and it is a recurring cause of agency enforcement actions.

That one rule constrains the agency payments stack in four ways most processors handle badly:

• Settlement destination. Settlement must land in the fiduciary account while commission and fee revenue belongs in operating. A single-account processor forces a manual sweep, and manual sweeps are where commingling findings come from.
• Net versus gross settlement. A processor that deducts its fees before depositing has taken money out of a fiduciary account to pay an agency expense. Gross settlement with separately invoiced fees is the clean structure, and it must be asked for, because netting is the default.
• Chargebacks and returns. A reversed premium payment debits the account it settled into, so the carrier's money covers the agency's dispute until it is made good.
• Reserves and holds. Any provider that reserves against future liability by holding back settlement is holding back fiduciary funds.

That last point is the strongest argument against running agency premium collection through a payment facilitator that boards sub-merchants under its own master account and sets reserves at its own discretion. Stripe and Square are excellent products that are structurally awkward here for exactly that reason: right for an insurtech MGA building its own billing engine and settling into accounts under its own control, and a poor default for an agency collecting premium it does not own.

Claims disbursement is a payouts problem, and the rail is the service level

Outbound is where insurers still lose the most goodwill, because the default rail is the slowest one available. The paper draft persists for real reasons: it handles multi-party payees natively, it asks nothing of the claimant beyond an address, and it is the only instrument here the insurer can stop before it is negotiated. It also produces the industry's least visible cost — uncashed items, which become unclaimed property subject to state escheatment rules.

Two design rules follow. Offer the claimant the choice and make the slow rail the fallback rather than the default — a claimant who selects instant payment to a debit card has also just supplied a verified identifier, which is worth something to the fraud team. And remember that instant rails are irrevocable: RTP and FedNow are credit-push and final on acceptance, so a claim payment sent to a fraudulently substituted account cannot be recalled as of right. Account ownership validation and out-of-band confirmation of any mid-claim change of bank details have to exist before speed is switched on, not after the first loss.

Multi-party payees are why paper survives: a property loss payable jointly to the insured and a mortgagee, or an auto claim payable to the insured, the repair facility and a lienholder, has no clean electronic equivalent on most rails. InvoiceCloud describes claim disbursement by direct deposit, push-to-debit and virtual card with multi-party support, which addresses a real gap — and is worth testing against the carrier's actual payee patterns.

EBPP platforms versus general-purpose processors

Vendors selling into insurance fall into three groups with different jobs, and the buying error is treating one as a substitute for another.

Biller-direct EBPP platforms. InvoiceCloud and Paymentus sit between the policy administration system and the policyholder: they present the bill, take payment across web, mobile, IVR, text, agent-assisted, kiosk and walk-in channels, and post results back with reconciliation. What a carrier buys is the integration and the channel coverage — InvoiceCloud names Guidewire PolicyCenter, BillingCenter and ClaimCenter, Duck Creek and Sapiens among its connectors, removing a build that would otherwise dominate the project. The structural point is that these are not processors: Paymentus states in its 10-K that its services do not involve the movement of funds directly, and depends on third-party processors and sponsor banks for settlement, which is also why its gross margin runs near 25 percent. There is still an acquirer underneath.

Acquirers and merchant services providers. Global Payments and Elavon will underwrite the carrier or agency, hold the merchant agreement, settle funds and carry the acquiring risk. They will not supply policy-administration integration, lapse-aware dunning, per-state fee configuration or claims disbursement. Buying an acquirer and expecting the insurance workflow is the expensive version of this mistake, because the gap surfaces after the agreement is signed — and Elavon's published US Terms of Service set an initial three-year term renewing automatically for successive two-year terms, with funds required to remain available in the designated account for at least 180 days after termination.

Rail and workflow specialists. GoCardless is the strongest option here for mandate-based premium collection in the UK and Europe, and a poor one for anything payable at the moment of sale, since a mandate takes days to establish. BILL is an accounts payable platform: relevant to paying adjusters and restoration contractors, irrelevant to premium.

What to settle before signing anything

These questions separate a workable insurance payments stack from one that fails its first audit. None is about price.

1. Which flows is this vendor actually for? Get inbound and outbound answered separately, in writing. A vendor excellent inbound and improvising outbound is a fine choice if the carrier knows that going in.
2. What is the system of record, and who writes back to it? If the platform cannot post a payment, a partial payment, a reversal and a refund back automatically, someone is doing it by hand — and that person is the reconciliation risk.
3. Is the fee-bearer decision configurable by state, and who signs off on its legality? The vendor will configure it. The vendor will not indemnify the carrier for having configured it wrongly.
4. Does settlement land gross or net, and in which account? On the agency side this is the fiduciary question, and it has one acceptable answer.
5. Does the dunning sequence take its schedule from the cancellation notice rules? Ask to see the state configuration, not the feature list.
6. What happens on exit? Who owns the stored payment credentials, and can they move without asking every policyholder to re-enter a card. If they cannot, the switching cost is the whole book.

Frequently asked questions