Award Category: Best learning technologies project - commercial sector (International) --

How can L&D scale employees’ language skills assessment without sacrificing trust, validity or learner experience? In this session, EPAM shares how we redesigned enterprise English speaking assessment using a human-in-the-loop model: an AI-led CEFR-aligned interview paired with certified human scoring and QA. You’ll see the end-to-end journey (test taker → assessor → results → learning pathway), the governance that protects reliability, and the change-management approach that helped drive adoption. We’ll unpack measurable outcomes—reducing turnaround from ~21 days to 2–7 days, lowering cost per test by 40%, and achieving strong user satisfaction — plus the practical lessons for teams considering AI-enhanced assessment in high-stakes corporate environments.

• By the end of this session, participants will be able to:
• Select and justify a human-in-the-loop assessment model (what AI does vs what humans do) for a scalable, high-stakes corporate context.
• Draft a phased rollout plan for an AI-enabled assessment (pilot → launch), including stakeholder messaging, onboarding and support/escalation routes.
• Build a reliability and governance checklist (standards alignment, calibration, QA, monitoring, re-test/appeals) to protect validity and trust.
• Define and calculate a compact impact scorecard (turnaround, cost per test, capacity uplift, user satisfaction) and translate it into a story suitable for business stakeholders.