Credicorp uses machine learning and AI models to support lending decisions. Here is how we approach fairness and explainability in AI-driven credit assessment.
What AI does in our decision process
Our AI models analyse patterns in company financial data — primarily bank transaction history from Open Banking and business credit data — to assess the likelihood that a company will repay a facility. The model outputs a risk assessment that informs (but does not automatically determine) the credit decision. A human confirms every payout — the final decision is not fully automated.
Protected characteristics and lending decisions
Our AI models are trained on financial data — revenue, cashflow, credit history, and business performance. They do not use protected personal characteristics (such as race, religion, gender, age, or disability status) as inputs. Lending decisions must not discriminate on these grounds, and we monitor our model outputs for differential impact across groups.
Regular model review
AI models can develop biases over time as the world changes and the data they were trained on becomes less representative. We review our models regularly and assess their performance and any differential outcomes across customer segments. Where bias is found, we address it — either by retraining the model or by applying adjustments to its outputs.
The right to request a human review
If you believe our decision does not reflect your company's true position, you can request that a member of our credit team reviews your application manually. See can I ask for a human to review my decision? for how to do this.
Explainability
We can provide a general explanation of the factors that influenced your decision — for example, whether cashflow patterns, director credit history, or trading profile were material. We cannot share model weights or proprietary code. If you receive a decline and want to understand the key reasons, contact us through the General Support Enquiry form.
Ongoing regulatory landscape
The UK's approach to AI regulation in financial services is evolving. We monitor developments — including the FCA's guidance on AI in financial services and the broader UK AI regulation proposals — and will update our approach as requirements become clearer.
See also: How the AI decision engine works, Can I ask for a human to review my decision?, Is the decision fair and unbiased?.