Survey Shows Strong AI ROI but Leaders Lack Confidence in Decision Explainability

FICO’s new "State of Responsible AI 2026" survey reveals that a large majority of senior AI professionals believe their investments are delivering the promised financial returns, with 85.1% reporting ROI at or above initial expectations. Despite this success, AI currently powers just 28.5% of decisions that directly affect customers, and only 5.2% feel highly capable of justifying those decisions to regulators or end users.

Executives point to difficulties in articulating why a particular model output was chosen, especially when data sources and model architecture must be traced for credit or fraud cases. While 82.8% consider their firms ready for forthcoming regulations, responsible-AI practices rank lowest among operational standards. The survey notes that most companies rely on disparate platforms, hindering consistent governance, and that a shared deployment platform is seen as a major ROI opportunity, though only 8.2% have fully implemented one. Early exploration of agentic AI is common, but security and data-integration concerns remain significant barriers.