How AI-Augmented Training Improves Worker Productivity
Artificial intelligence can boost productivity not only where tools are used directly in day-to-day work. The newly released ai:conomics paper examines another channel: AI-supported training as an indirect lever for performance gains—even for employees who do not use AI themselves.
A growing body of research already shows productivity gains from AI (Brynjolfsson et al., 2025; Dillon et al., 2025; Dell’Acqua et al., 2025), with most focusing on direct effects—for example, when employees use general-purpose AI tools like ChatGPT or Microsoft Copilot to support their work tasks. Beyond that, there are indirect effects that also affect employees who do not use AI. ai:conomics investigates this indirect effect in a real-world environment: an empirical study in a call center of a large European financial services firm. There, a tool was rolled out step by step that provides individualized performance feedback on calls, serving as a basis for more targeted training oft he call center agents. The staggered rollout of the AI application enables clean identification of causal effects.
The study shows: Faster call handling overall—with different drivers for short-tenure and long-tenure employees.
AI-augmented training reduces handling time per call by an average of 10%. The productivity increase is larger for short-tenure employees (17%) than for long-tenure employees (7%).
The reasons for these gains differ by group: Less-experienced employees close knowledge gaps, make fewer mistakes, and put callers on hold less often to consult colleagues. This suggests that AI-augmented training moves them up the experience curve faster and speeds up call processing. For long-tenure employees, the training mainly improves communication style: conversations become more precise, filler words and hedge words decrease—talk time shortens and calls become more efficient.
Overall, the study shows that AI-augmented training contributed to a reduction in productivity differences across employees and that efficiency gains go hand in hand with improved customer satisfaction.
Why does this matter?
The findings show that AI-based feedback in training leads to measurable changes in work performance—through different levers for different experience levels. In this way, AI raises team performance via the training channel—even among employees who do not use AI themselves
