Case study · August 1, 2026 · 7 min read

The AI amplifier loop: how people create value beyond the baseline

A field study suggests that the advantage is not simply using AI. It is directing the work, applying judgment, and improving the result.

An illuminated interface representing a human approval step in an AI workflow

The human contribution becomes clearer when the workflow shows where direction and judgment enter.

The case study

KPMG and the University of Texas at Austin studied 523 US-based early-career professionals working with an AI agent on tasks that mirrored real client work. Researchers first established an AI-only baseline, then compared how people collaborated with the same kind of system.

The result was not explained by traditional measures of knowledge or capability. The difference came from how people directed the AI, evaluated its work, and refined the result.

Three collaboration profiles in the study
AI Amplifiers50.1%Outperformed the AI-only baseline through orchestration, domain framing, and refinement.
AI Delegators25.8%Produced results comparable to AI alone by accepting competent output with little scrutiny.
AI Apprentices24.1%Had strong foundations but critiques often failed to steer the work toward a better result.

The amplifier loop

The useful response is not to ask people to “use AI more.” It is to teach a repeatable loop that makes human value visible:

FrameDefine the outcome, audience, constraints, and decision.
DirectGive the AI a focused role, useful context, and a clear next move.
CritiqueCheck evidence, reasoning, gaps, and whether the work answers the real question.
RefineUse the critique to improve the output, not just to decorate the review.

Copy this coaching brief

Task: [what decision or outcome matters?] Frame: [audience, constraints, quality bar] Direct: [role, context, sources, first move] Critique: [what evidence, assumptions, gaps, or risks need checking?] Refine: [what should change in the next round?] Human value: [where did judgment, expertise, or accountability improve the result?]

A small team application

A product team uses the loop for a weekly customer-signal report. The assistant clusters feedback and drafts themes. The product manager frames the decision the report must support, directs the assistant toward the approved sources, critiques unsupported claims and missing segments, then asks for a refined version with confidence notes and open questions.

That workflow creates a better report and a better learning signal. The team can see which parts AI handles well, where domain expertise changes the result, and what coaching future users need.

The takeaway

The strongest AI users are not merely prompt writers or final approvers. They are workflow designers. They help AI aim at the right problem, test the output against reality, and improve it until it is ready for a human decision.