Using AI across more credit workflows does not fully explain which organizations report the highest impact. The Global Institute of Credit Professionals (GICP), a Fitch Learning-owned professional body for credit professionals, found AI governance maturity and changes to work rose alongside reported AI impact in its global “Credit in the Age of AI” research.
GICP found almost half of organizations using AI across 25% to 49% of workflows reported high organizational impact, rising to 66% among those using it across 50% to 75%. The higher-adoption result needs caution: the report’s chart includes 50 respondents in the 25% to 49% group but only six in the 50% to 75% group and one above 75%. GICP itself says the high-adoption data should be interpreted cautiously.
Governance maturity and workforce evolution align with success
GICP constructed a Governance Maturity Index around accountability, oversight and model-assurance practices. Among organizations with eight or nine governance factors in place, 67% reported high AI impact, compared with 23% of those with zero or one.
Organizations reporting low AI impact had made just over one workforce or workflow change per organization on average, compared with nearly two among medium-impact organizations and about 2.6 among high-impact organizations. The changes measured included job responsibilities, workflows, hiring criteria and skill requirements.
Core credit decisions yield the highest reported impact
The strongest reported impact also appeared closest to credit decisions. Risk assessment, scoring and early warning received the highest average organizational-impact score in the study, at 3.16 on a five-point scale. GICP found higher uptake in easier-to-check activities such as drafting, analysis and research, while decision-centric uses placed greater demands on validation, monitoring and control.
The operating gap is also visible in investment priorities. GICP said 65% of respondents cited at least one leadership, strategy or governance-related barrier, while 38% were investing in governance, guardrails or AI strategy.
Risk management and regulatory compliance constrain scaling
Earlier credit-risk research identified a similar implementation constraint. McKinsey surveyed senior credit-risk executives at 24 financial institutions in 2024 and found 75% identified risk and governance as significant barriers to scaling generative AI. Only one-third of the institutions had established a center of excellence to manage generative AI use cases.
For U.S. creditors subject to Regulation B, using AI in credit decisions does not change adverse-action explanation requirements. CFPB guidance says creditors using complex algorithms, including AI or machine learning, must still be able to provide specific and accurate principal reasons for adverse action, while the regulation’s official interpretation says disclosed reasons must accurately describe factors actually considered or scored by the creditor.
The National Institute of Standards and Technology’s voluntary AI Risk Management Framework separately sets out governance practices covering defined responsibilities, staff training, human oversight and AI testing.