KPMG LLP and the McCombs School of Business at the University of Texas at Austin published a field study of 523 US-based early-career professionals working with AI agents, released July 23, 2026 and appearing in Harvard Business Review. Participants completed tasks mirroring real client work using an AI agent, and results were compared against an AI-only baseline established by having agents complete the same work unassisted. The analysis identified three performance profiles: AI Amplifiers (50.1%) exceeded the AI-only baseline, AI Delegators (25.8%) matched it, and a third group fell below. The authors conclude that professionals with nearly identical knowledge and skill can produce very different outcomes once AI enters the workflow, with performance depending on how they direct the AI and evaluate its output, and that many employees who do not yet exceed AI could improve with support.