GICP study links AI governance to credit sector outcomes
The Global Institute of Credit Professionals (GICP) has published research showing that AI governance maturity, rather than the volume of AI deployment, is the primary differentiator between credit organisations achieving meaningful business outcomes and those that are not. The report, "Credit in the Age of AI", surveyed organisations across the credit lifecycle and found that 67% of those with eight or nine governance measures in place reported high AI-enabled impact, against just 23% of those with zero or one measure.
The figures point to a structural gap in how the industry is approaching implementation. Fifty-six per cent of surveyed organisations currently use AI in fewer than 25% of their credit processes, and 14% have not deployed AI at all. Adoption is concentrated in lower-risk tasks such as document drafting, financial analysis and market research. The areas where AI is reportedly delivering the greatest impact, namely risk assessment and credit scoring, remain less penetrated.
The governance paradox
The report identifies a telling disconnect between where organisations recognise barriers and where they are directing investment. Sixty-five per cent of respondents named leadership and governance factors as the single biggest obstacle to successful AI adoption. Yet only 38% plan to prioritise spending in that area, suggesting that budget allocation continues to lag behind diagnostic awareness.
Organisations reporting high AI impact also introduced workforce and workflow changes at more than twice the rate of lower-impact peers, averaging 2.58 organisational changes compared with 1.22. The finding reinforces an argument that has become common in enterprise transformation literature: technology alone does not drive outcomes; it is the surrounding operating model that determines whether capability translates into performance.
Andreas Karaiskos, Executive Director of the GICP, said: "The organisations seeing the greatest impact are investing in governance, workforce capability and operating models that enable AI to be embedded effectively into day-to-day credit decision-making."
Two waves, one readiness problem
The GICP frames the current moment as the transition between two distinct phases of AI value creation. The first wave centred on productivity gains from drafting, summarisation and research tools. The second, now emerging, involves applying AI to underwriting, credit assessment and risk scoring, activities where the commercial and regulatory stakes are considerably higher.
The scale of the gap between aspiration and readiness is significant: among organisations that expect disruption from generative AI, autonomous decision-making or real-time monitoring, 79% are not yet deploying AI in the relevant functions.
This readiness deficit is relevant beyond the credit sector. Regulators in the UK and EU have increasingly focused on model explainability and auditability in high-stakes financial decisions. The EU AI Act classifies credit scoring systems as high-risk AI, triggering conformity assessment, transparency and human-oversight obligations that will phase in progressively through 2026 and 2027. In the UK, the FCA has signalled continued scrutiny of algorithmic decision-making in consumer credit under its Consumer Duty framework. Organisations moving into the second wave of AI adoption, particularly those using models in underwriting and scoring, will need governance infrastructure not merely as a performance lever but as a compliance prerequisite.
The GICP's findings add quantitative weight to an argument that enterprise software buyers and AI vendors alike have been making: that deployment counts matter less than the institutional fabric around them. For credit organisations still in the shallow end of adoption, the study provides a clear, if uncomfortable, read: the governance investment gap is also an outcomes gap.