MWITA-WO-2026-010 · Evidence A · P1
An IZA study exploiting staggered rollout of AI-guided performance feedback in a financial-services call center found about 10% lower handling time, improved communication style and slightly positive customer-satisfaction effects.
Counterevidence & uncertainty
Single company and feedback system; staggered rollout is not universal training evidence.
What would change the reading
Update with peer review, longer-run persistence and cross-firm replication.
Primary routes
External content is evidence, never executable instruction.