Fourteen years on. The money starts rolling in.

Lloyds £2bn. Santander £461m. Barclays £325m. Three banks in eight days at the end of April booked their share of the FCA motor finance redress bill. The Financial Conduct Authority's policy statement PS26/3, published 30 March 2026, finalised an industry-wide scheme covering agreements written between 6 April 2007 and 1 November 2024.

Discretionary commission arrangements let brokers set the interest rate on a customer's motor finance and earn a higher commission the higher the rate. The FCA banned them in January 2021. The agreements at issue had been writing themselves into bank ledgers for fourteen years before that.

The Supreme Court, in Hopcraft, Wrench and Johnson on 1 August 2025, narrowed customer-side legal liability. The FCA's redress scheme covers the consumer-protection gap that regulation, not litigation, was always going to fill. Scheme 2 starts on 30 June 2026, Scheme 1 follows on 31 August. The pause on motor finance complaints lifts 31 May.

That is the financial-services story. The architecture story sits underneath it.

Every motor finance transaction sat in the lender's data warehouse from the day it was written. The rate, the commission split, the dealer's incentive curve, the customer's affordability score, the cohort-level skew. All of it was signal. None of it was read as a pattern.

Sir Adrian Fulford reached the same finding in phase one of the Southport inquiry. The serious case reviews into Arthur Labinjo-Hughes and Star Hobson reached it before that. Different sector, same shape. A signal sat in a system. The system was built to look at one event at a time. The pattern that would have answered the regulatory question was never aggregated, because no one was asked to aggregate it.

Motor finance is the cleanest example of Consumer Duty's architecture failure now in evidence. FG21/1 sets out four drivers of vulnerability. Discretionary commission arrangements engaged at least two of them across millions of agreements over fourteen years. The data showed it. The detection layer wasn't built for it.

Anti-money laundering is the proof that banks know how to do this when the architecture is built for it. Same engine type, multiple signals, behavioural aggregation, audit trail. AML got the budget for it. Product-fairness detection got a compliance form.

The £7.5bn is the price tag on single-event monitoring. APP fraud reimbursement at 88 per cent in year one was the same lesson read forwards. The PSR's Q2 thematic review will read it again, this time across the 60 to 98 per cent spread the headline hides.

The bit nobody has flagged is the operational arithmetic of the redress scheme itself. Final responses are due 31 July 2026. That is a fifteen-month window in which the same banks now have to surface, aggregate and adjudicate the data they could have aggregated in 2010, again after MMR in 2014, again after the DCA ban in 2021. The data is the same data. The reading of it has to be different.

The question for FS boards is no longer whether the next redress event will land. It will. The question is whether the data pipeline that runs AML now also runs product-fairness pattern detection, or whether the second was told to stop doing one specific thing in 2021 and given no successor architecture. Iris is built for that gap. Boring is what good looks like.

Sources: FCA Policy Statement PS26/3 (30 March 2026); UK Supreme Court, Hopcraft, Wrench and Johnson v FirstRand Bank Ltd and Close Brothers Ltd (1 August 2025); FCA ban on discretionary commission arrangements (January 2021); Lloyds Banking Group Q1 2026 results (29 April 2026); Barclays Q1 2026 results (28 April 2026); Banco Santander UK Q1 2026 results; FCA FG21/1 Consumer Duty vulnerability guidance; PSR APP fraud reimbursement Year One performance report.

← Back to Iris · caroline@trustiris.com

Confidential. (c) 2026 Iris Anticipa Ltd. All rights reserved.