Most companies suffer some risk-related financial loss deploying AI (www.reuters.com)

An EY survey of 975 executives at companies with annual sales over $1 billion (July–August 2025) found that nearly every large firm that has rolled out AI experienced some initial financial loss, with combined losses estimated at $4.4 billion. Common root causes were compliance failures, flawed or biased model outputs, and unintended disruptions to sustainability goals; reputational damage and legal problems were reported less often. Key business metrics—revenue growth, cost savings and employee satisfaction—also trailed expectations, even as most firms remained optimistic that AI will deliver long‑term productivity gains. EY framed the work through a “Responsible AI” lens, measuring governance, usage guidelines and compliance monitoring. For the AI/ML community this underscores that deployment risk and value capture are distinct challenges: model performance alone won’t guarantee business benefit without governance, observability and integration into operational processes. Firms reporting more mature Responsible AI policies saw better sales, cost and employee outcomes, implying that investment in model validation, bias mitigation, regulatory compliance, human‑in‑the‑loop controls and monitoring pays off. The findings signal demand for tooling and practices around robust testing, explainability, compliance auditing and lifecycle management to reduce early losses and convert productivity improvements into measurable business value.