Consumer credit in Europe is entering a more demanding phase. Regulators are making it clear that responsible lending can no longer be treated as a slogan: it has to be something lenders can evidence and defend, consistently, across products, channels, and markets.
Customer expectations haven’t softened in response. Credit decisions still need to be fast, digital, and frictionless, even as affordability gets harder to assess in a world of variable income, multiple concurrent credit products, and rising payment stress.
The way through is better data and better decisioning, backed by scalable loan management software and lending analytics that can adapt as regulation and borrower behaviour keep changing.
Regulators now expect proof, not good faith
The driver is the updated EU Consumer Credit Directive (Directive (EU) 2023/2225), which replaces the 2008 framework and brings new products into scope for the first time. The headline focus is affordability and over-indebtedness, but the deeper change is about proof.
As Zitah McMillan, CEO of Finexos and a former Executive at the FCA, puts it: “The biggest change is how you explain the process you went through and the decisions you took.”
Historically, lenders could demonstrate compliance by showing that creditworthiness checks had been performed. That’s no longer enough. “Previously, we could say we checked their creditworthiness,” McMillan says. “Now we need to show whether the loan was actually affordable to that person.”
Creditworthiness and affordability are no longer treated as the same thing. Regulators expect lenders to show that a loan made sense in the context of the borrower’s real circumstances. As McMillan says: “It’s gone from saying ‘I can lend in good faith’ to having to prove it.”
2026 is the year of execution, not interpretation
EU member states had to transpose the directive into national law by November 2025, with the new requirements applying from November 2026. That makes the coming year about execution: turning policy language into operational reality across decisioning engines, customer journeys, servicing processes, and loan management systems.
Manual workarounds and informal judgement won’t hold up under scrutiny. Consistency, repeatability, and evidence matter more than good intentions.
One directive, applied differently market by market
The directive aims to harmonise consumer credit standards across the EU, but national regulators can still apply and supervise the rules differently. In markets without a single conduct regulator, the regulators themselves are still adapting to the expanded scope.
For lenders operating across multiple markets, that creates a familiar problem: keeping a consistent affordability framework while flexing for local expectations. Modular lending platforms and loan analytics software help here, letting policy updates and decisioning changes roll out centrally while still accommodating local nuance.
Why affordability models are being re-examined
Affordability is no longer a clean calculation. Borrowers today often juggle instalment loans, revolving credit, overdrafts, BNPL arrangements, subscriptions, and informal obligations, while income can be variable or seasonal. Two applicants with identical salaries can have very different cash-flow realities.
Many lenders, McMillan notes, either didn’t explicitly assess affordability in the past or built models years ago that have since drifted. “Over time, new people join organisations, thresholds get tweaked, and things drift,” she says. “You have to go back to first principles.”
The regulatory change forces lenders to look again, without leaning on the assumption that their existing process is already good. “It won’t be the case that everyone has to start from scratch,” she explains. “But everyone has to think about the legislation in a way that isn’t shaped by what they’ve done previously. You have to look at what the regulator is looking for now.”
From approval to justification
The question in 2026 isn’t “can we approve this loan?” It’s “can we justify approving this loan?”
That justification has to link policy, data, and decisioning logic in a way that can be reconstructed later by regulators, auditors, boards, or funding partners. Affordability has become as much about governance and evidence as it is about risk.
Data is only useful if you can defend it
Most lenders don’t lack data. They lack clarity about which data actually improves decisions.
Affordability assessments typically combine bureau data, self-declared information, bank-transaction data, internal performance history, and third-party sources. Each input has value, but none should be accepted at face value. “I don’t think lenders ever take anything at face value,” McMillan says. “When money is involved, face value is not a good part of your decisioning.”
Every data source should be tested against outcomes and checked against business rules. “Over time, what’s happened is data ingestion creeps up,” she explains. “More layers of data come in, often from third parties, and no one’s checked recently whether they’re actually helping.”
The regulatory transition is a natural point to ask: is this input helping make better decisions? Could I rely on it if I had to submit it to a regulator? Does it lead to good customer and business outcomes? Many lenders are now back-testing affordability assumptions against real slices of their loan books and adjusting thresholds based on actual performance rather than theory. AI-enabled lending analytics can speed this up, particularly where internal teams are stretched.
Explainability is still being worked out
Explainability and reporting remain uncertain: local regulators will determine formats, metrics, and frequency over time. McMillan describes the moment with a metaphor: “We’re in a Rubik’s Cube moment. We don’t yet have one side fully coloured in. We’re trying to figure out which side we’ll get first, what colour it will be, and how we solve the puzzle from there.”
Uncertainty doesn’t mean inaction. Lenders can start now with decision logs that show how over-indebtedness has been actively prevented, including loans not made or amounts reduced because of affordability assessments.
AI raises the stakes on accountability
As AI becomes more embedded in credit decisioning, governance isn’t optional. Under the Consumer Credit Directive and the EU AI Act, lenders must be able to demonstrate explainability, accountability, and the possibility of human intervention.
Boards feel this acutely. “There’s a fear that if you can’t explain what the AI did, you’re on the hook for it,” McMillan explains. “You can’t say ‘the AI did it’: nobody will accept that.”
There’s a pull in the other direction too. “There’s also the fear of missing out,” she says. “Investors expect to see AI. If it’s not in your updates, they think you’re stuck in the mud.” That tension, between caution and expectation, is one responsible AI providers need to understand and address. As McMillan puts it: “Boards don’t need to understand the code, but they do need to understand what the AI is doing and how it’s doing it.”
The mindset shift for 2026
For McMillan, the most important change ahead is philosophical: “It won’t be ‘can I make money on this loan?’ It will be ‘can the borrower afford this loan?’, and then, ‘can I lend it profitably?’”
With the regulatory clock running, she warns against waiting: “You don’t know when your regulator will come knocking. In some markets, it may be sooner rather than later. So use this time to get into a really good place.”
Closing thoughts
Affordability and over-indebtedness are no longer side themes in consumer credit. They’re the lens regulators, boards, investors, and customers now use to judge lenders. In 2026, better outcomes will come from better data and better decisioning: lending analytics and loan management software built to stand up to scrutiny, not just process applications.
CreditOnline’s Loan Management System is built around exactly that kind of evidence trail, logging the data and decisioning logic behind every approval so it can be reconstructed later for a regulator, auditor, or board.