Executive Summary

Consumer credit becomes a technology infrastructure problem when a lender has to decide whether a customer can afford a specific agreement while that customer is still at checkout. The UK's Deferred Payment Credit regime makes that tension concrete. Lenders need data they can use at the point of decision, a way to vary the depth of the assessment with the risk, and payment systems that can carry the decision through to repayment. A retailer can change the appearance of a payment option in a release. The lender's ability to make and explain the underlying decision takes longer to build.

The signal: a credit decision at the point of purchase

The UK's Financial Conduct Authority (FCA) published Policy Statement PS26/1 on 11 February 2026. The regime began on 15 July 2026. The statement concerns Deferred Payment Credit (DPC), a form of interest-free finance for purchases, repayable in no more than 12 instalments over no more than 12 months. The regulatory change covers in-scope third-party lending. The product is often marketed as buy now, pay later, a label that also covers arrangements outside this particular change.

The operative detail is in the FCA's final policy statement and rules. Lenders must carry out a proportionate creditworthiness assessment before entering into each in-scope DPC agreement. That includes small agreements below £50, as A&O Shearman's reading of the final rules also explains. The FCA allows for differences in the depth of the assessment. A returning customer may already have relevant information on file, and the lender still needs to decide whether it is sufficient for the new agreement. This puts a fresh decision at the checkout even where some of the evidence can be reused.

The FCA's product information rules also require lenders to say whether they will obtain information from a credit reference agency if they already know, or whether they may obtain it if the decision is still open. That distinction matters at checkout. The borrower sees the disclosure before the agreement, and the lender may still be deciding which evidence the assessment needs.

Where the old lending process starts to strain

Credit decisioning is the process of turning information about a borrower and a proposed agreement into a lending decision. Embedded finance places that process inside another service, such as a retailer's checkout. The journal's analysis of the infrastructure behind digital consumer credit follows the decision inside that payment experience. The broader problem starts when systems built for separate applications, periodic data updates and later reconciliation are asked to operate within the same purchase.

Advertise with us

Consider a hypothetical shopper who returns for a second purchase before paying off the first. The merchant knows the new basket total. The lender's servicing system knows the remaining instalments. A payment processor may know that the most recent collection failed before that information appears in the lender's decisioning system. The separate records can all be accurate at the time they were generated. The decision can still be poor if the system treats them as equally current. A slower checkout buys little if the assessment is still using stale information.

The issue is also what the records can show. Regular obligations may be visible in one data source and missing from another, while consumers with short credit histories may leave lenders with limited evidence. The journal's analysis of why rent payments are often missing from credit data illustrates the gap. A missing record can reflect the way information is collected. For a lender, the practical question is how to distinguish that gap from a verified absence of obligations and when the uncertainty warrants more information.

Why the decision and the payment have to stay connected

Checkout is where the amount and terms of the prospective agreement become concrete. If the basket changes, the lender needs a reliable way to tie the decision to the agreement the customer actually accepts. After purchase, the payment and servicing systems need to record what happened to that agreement. In a hypothetical case, a partial refund changes what is owed, and if the refund reaches the merchant's system before the repayment schedule changes, the lender can collect the wrong amount or carry the wrong exposure into a subsequent assessment.

This is the wider context for Philip Belamant's work on consumer-credit infrastructure. A decision engine can respond quickly and still depend on fragmented records. A payments system can collect instalments reliably and still fail to return the information the next decision needs. The advantage comes from connecting those systems so a lender can reconstruct what it knew at the time, which terms the customer accepted and what subsequently happened to the balance.

What to watch for lenders and consumers

For lenders, checkout conversion is only one measure of whether the system works. Alongside it, operators can track how often a decision needs additional evidence, how long those cases take and what happens to repayments after approval. Where assessments rely on older information, reviewing complaints and missed payments can help identify where an apparently efficient process is producing weak decisions. An approval rate on its own cannot show whether faster decisions improved access for people who could afford the agreement.

For consumers, the potential gain is an assessment that can account for relevant circumstances with less repeated form-filling. The risk is that a lender turns uncertainty into a blanket decline. The journal's reporting on access for first-time cardholders in Mexico offers a parallel about serving people with limited borrowing histories. The UK rule applies to the DPC agreements within its scope. What travels across markets is the question of whether a lender's systems can assess a borrower when conventional records are sparse.

The Federal Reserve's 7 October 2026 consumer credit release supplies wider US market context. Its aggregate balances are a different measure from the performance of UK checkout lending. The FCA policy statement identifies an operating requirement. Evidence of its effects would come from outcomes such as approvals, repayment difficulties and complaints under the new regime.

Key Takeaways

  • The UK regime makes proportionate creditworthiness assessment part of each in-scope Deferred Payment Credit agreement.
  • A checkout decision is only as reliable as the systems that reconcile the proposed agreement with the evidence available at the time.
  • Faster decisions are valuable when lenders can also explain them and track what happens to borrowers after approval.

Conclusion

The Diaspora Operators series follows businesses that compete by changing the systems through which services reach customers. In consumer credit, the FCA policy statement makes one part of that thesis testable. A lender has to assess the agreement in front of it, then let servicing and repayment inform what happens the next time the customer arrives. The competitive question is whether it can do that at checkout speed while preserving enough evidence to explain the decision afterwards.