The Redesign Dividend: Why Most Organizations Still Aren't Getting Value From AI, and How Leaders Can Redesign Work for a Human-Agent Workforce

Abstract: Despite several years of heavy investment in generative and agentic artificial intelligence, most organizations report little enterprise-level value. Recent surveys of chief information officers and chief executives show that only a small minority of enterprises have realized significant or transformational value, that most AI use cases fail, and that many leaders struggle to demonstrate financial returns. This article argues that the gap is primarily a work design problem rather than a technology problem. Drawing on field experiments, sociotechnical systems theory, job design research, and economic studies of automation, it explains why tool-centric pilots and headcount-first strategies tend to disappoint. It then outlines five evidence-based responses: redesigning workflows end to end, writing explicit role-redesign plans for a human-agent workforce, installing stopping rules for low-value initiatives, reinvesting freed capacity, and governing AI economics through outcome measures. Organizational examples from financial services, retail, biotechnology, education technology, and software illustrate both promising practice and costly missteps. The article closes with three long-term capabilities that allow organizations to capture what it calls the redesign dividend.