Experiments & interventions
Small tests that solved real operational problems
Stroop-based attention training for issue recognition.
Agents were answering the surface question and missing the real issue. I used the Stroop effect as a diagnostic, banded agents by cognitive flexibility, and ran a three-week conditioning programme. 100% of lowest-band agents moved up at least one analytical tier; 50% advanced a tier after phase two.
See it inside the case study →Real-time monitoring before the bad rating lands.
Quality feedback arrived too late. I designed scheduled real-time review windows, intervention protocols and agent self-flagging. Monitored days hit 98–99% CSAT and the client doubled team headcount within three months.
Customer vocab playbook + service mechanics blueprint.
Customers describe the same issue in a dozen ways, and tenured agents were skipping due diligence. I built a playbook of real customer vocabulary and a decision-tree blueprint for every common process. Result: 44% more agents hitting perfect internal quality scores and a 135% productivity lift in six weeks.
QA frameworks, standards and review processes I operationalised.
Sanitised samples of real QA documentation: a weekly rolling QA model that feeds DSAT signals into scores, rating/coaching standards from a 30-day calibration sprint, and the end-to-end quality review workflow.