# How a lender assesses whether a company can afford to repay

Affordability is the question every responsible lending decision turns on: can this company comfortably repay this amount, over this term, out of the cash it actually generates? It is a different question from “will the loan be approved” — it is the work that sits underneath the answer. This piece sets out how that assessment is done for an incorporated business: the data a lender reads, the track record it weighs, the role people play alongside the algorithms, and the right you have under UK data-protection law to a human looking at the decision.

## The data a lender reads

A company affordability assessment is built from a handful of sources, each answering a different part of the question.
- **Bank statements.** The company’s account is the single most honest record of how money actually moves — income in, costs out, the rhythm of the trading month and how tight or comfortable the balance runs. It shows behaviour, not just position.
- **Open Banking.** With the company’s consent, Open Banking lets a lender read recent transaction data directly and securely, rather than relying on documents that may be weeks old. It is read-access to the same information, faster and harder to misstate. The operator explains what is and is not read in its note on [how Open Banking is used](https://credicorp.co.uk/news/open-banking-how-we-use-it/).
- **Business credit data.** The company’s bureau profile and public record — filed accounts, payment behaviour, any adverse markers — give an outside view of how the company meets its obligations. How that score is built is covered in [understanding business credit scores](/articles/understanding-business-credit-scores).
- **Track record.** How long the company has traded, whether its income is steady or lumpy, and whether it has handled obligations before all shade the picture. A thin file on a young company is read in context, not as a failure.

## From data to a view: what affordability really tests

The inputs feed one judgement: does the company generate enough cash, reliably enough, to meet the repayments without being pushed into difficulty? That is not the same as “does it have the money today”. It is about whether the repayments sit comfortably within the company’s normal cash flow over the term — with room for the lumps and gaps that real trading brings. A loan that the company could only repay if everything goes perfectly is not affordable; one that fits inside its ordinary rhythm is. The [true cost of the loan](/articles/the-true-cost-of-a-short-term-loan), line by line, is part of that test: the repayment has to be sized to the cash, not the other way round.

## People, not just algorithms

Models are good at sorting and flagging, and they do a lot of the first-pass work — checking the company exists, reading the statements, surfacing the patterns. But a model is a starting point, not the final word. Real trading is messy: a one-off large payment, a seasonal dip, a recent change of customer mix can all look like risk to an algorithm and turn out to be perfectly explicable. That is why the operator’s position is that decisions are reviewed by people, not settled by a score alone — the approach it describes as [affordability over algorithms](https://credicorp.co.uk/news/affordability-over-algorithms/). A person can read context an algorithm cannot, and can say yes to a company a crude model would reject, or no to one it would wave through.

## Your right to human review — Article 22

This is not only good practice; it connects to a right in UK data-protection law. Under Article 22 of the UK GDPR, an individual has the right not to be subject to a decision based *solely* on automated processing where it produces legal or similarly significant effects — and, where automated decision-making is used, to obtain human intervention, to express their point of view and to contest the decision. In plain terms: a meaningful decision should not be made by a machine on its own with no human able to look at it. A lending process that keeps people in the loop is consistent with that right rather than working around it. The operator sets out how this works in practice, including automated-decision review, on its [technology page](https://credicorp.co.uk/our-technology/).

## What this means for an applicant

For a director, the practical upshot is straightforward. The cleaner and more current the company’s data, the more accurately a lender can assess it — which is an argument for up-to-date filings, tidy bank conduct and, where offered, Open Banking access that lets the lender see the real position rather than guess at it. It also means an affordability decision is not a black box you have no recourse to: you can ask how it was reached and ask for a person to review it. A lender that assesses honestly is doing it to lend what the company can repay — which is in the borrower’s interest as much as the lender’s.

## The honest summary

A company affordability assessment reads bank statements, Open Banking data, business credit and track record to answer one question: can the company comfortably repay from the cash it generates? The data does the spadework, but people make the call — and UK GDPR Article 22 gives you the right to human involvement in a decision that significantly affects you. For the borrower, good data and an honest assessment point the same way: a loan sized to what the company can actually afford.

## Related
- [Understanding business credit scores](/articles/understanding-business-credit-scores)
- [The true cost of a short-term business loan, line by line](/articles/the-true-cost-of-a-short-term-loan)
- [When not to borrow: signs a short-term loan is the wrong answer](/articles/when-not-to-borrow)
- [Our technology, including automated-decision review (operator)](https://credicorp.co.uk/our-technology/)
- [Lending and regulation — Article 60B in plain English](/lending-and-regulation/)