New: the Kontor regulatory map, with UK, EU and US obligations mapped to controls and evidence. Read it

[ Blog / Platform ]

Banking as a service: what lenders actually need from it

The Kontor team·6 October 2026·4 min read

A fintech that wants to offer consumer loans typically reaches for a banking as a service platform expecting a complete solution. What it gets instead is a set of payment rails, a stored-value account and perhaps a card programme. The credit layer, the loan ledger, the interest accrual engine, the servicing workflows, the Consumer Duty disclosures, none of that comes in the box. The lender has to build it alongside, or bolt it on afterwards.

That gap is where lending programmes fail quietly. The payment infrastructure works. The credit infrastructure is improvised. This article describes what a BaaS platform must do to support a credit book properly, and where the line sits between banking infrastructure and lending infrastructure.

What a credit-grade BaaS platform actually consists of

A platform that is genuinely fit for consumer credit has several distinct layers. Payment rails are one of them, and not the hardest one. The components that matter for a lending programme are the ones that touch money movement, regulatory record-keeping and customer outcomes over the life of a loan.

  • Event-sourced loan ledger. Every transaction, a drawdown, a scheduled repayment, a late fee, a rate change, is recorded as an immutable event. The current balance is derived from the event log, not stored as a mutable field. This means you can reconstruct the account state at any past moment, which regulators and complaints handlers will eventually ask you to do.
  • Jurisdiction-aware rule engine. Consumer credit rules differ materially across markets. The UK Consumer Credit Act and CONC govern fee caps and arrears treatment differently from the EU Consumer Credit Directive or US Regulation Z. A platform that hardcodes one regime forces a rebuild when you add a second market. Jurisdiction packs bind the same engine to each market's rules without forking the codebase.
  • Auditable servicing layer. Every decision made during the life of a loan, a forbearance offer, a payment allocation, an arrears letter, must be traceable to a rule or an instruction. Agent Operating Procedures formalise what servicing agents are permitted to do and produce a record that compliance teams can inspect.
  • Dual-clock accrual. Loan schedules run on a calendar that does not match the system clock. A payment due on 29 February on a monthly schedule, an interest accrual over a bank holiday weekend, these require a ledger that separates business-date logic from wall-clock time. Getting this wrong silently misstates customer balances.
  • Portable credit book. A lender that cannot move its loan portfolio without a full system migration is operationally captive. An event-sourced ledger with clean interfaces makes migration without a big-bang cutover tractable.

How a credit book event flows through the platform in practice

Take a concrete case: a borrower's direct debit fails on the due date. The platform must do several things correctly and in sequence, not just log the failed payment and wait for someone to notice.

  1. Record the event, not just the outcome. The ledger appends a payment_failed event with the business date, the amount expected and the reason code from the payment scheme. The account balance does not change yet; the event log is the source of truth.
  2. Apply the jurisdiction's arrears rules. Under CONC 7.3, a lender must treat customers in financial difficulty fairly and cannot apply a default charge that is disproportionate. The rule engine reads the jurisdiction pack for the account, determines whether a fee is permissible and at what level, and appends a fee_applied event or suppresses the fee accordingly.
  3. Trigger the servicing workflow. The platform routes the account to the correct arrears queue, generates the required notice (in the UK, a default notice under section 87 of the Consumer Credit Act for regulated agreements), and logs which template was used and when it was sent.
  4. Create the audit record. Every step, the failed payment, the fee decision, the notice, is linked to the originating event. If the borrower later complains and the Financial Ombudsman Service requests the account history, the full sequence is retrievable without manual reconstruction.

What this gives a lender is a complaints and audit posture that does not depend on anyone remembering what happened. The record is the system's natural output, not something assembled after the fact.

Why payment rails alone do not make a lending platform

The objection lenders sometimes raise is that the payment infrastructure handles the hard part and the credit logic can be layered on top cheaply. In practice the opposite is true. Moving money is a solved problem across most banking as a service providers. Accruing interest correctly across a rate change mid-cycle, allocating a partial payment to principal, fees and interest in the right order under Consumer Duty, generating a statement that a regulator would accept as accurate: these are the problems that take engineering time and produce regulatory exposure when they go wrong.

Kontor is built as credit infrastructure first, with payment connectivity as one input to the ledger rather than the other way around. The distinction is not architectural preference; it is the difference between a platform that can be examined by a regulator and one that cannot.

A banking as a service platform earns the name in consumer lending when it can show, for any account and any moment in its history, exactly what happened, why and under which rule: payment rails are the entry point, not the substance.

[ More from the blog ]

Run your receivables on Kontor.

Bring a product spec or a portfolio extract. We'll show you how the ledger, decisions and agents would run it, and what your regulator would see.