A legacy menu-based IVR system stood in front of a high-volume contact center. The customer would navigate through menus led by operators, hang up and dial back, adding calls to the volume of work even as the IVR system failed to provide effective self-service.

Replace navigation by menus with intent-based conversation. The customer just tells the system what he wants, and the AI solves the simple requests and transfers the rest along with the whole conversation context. Transfer quality became the key KPI since bad transfers cost more than calls that AI won’t handle.

Integration into telephony before the legacy contact center platform, training of speech recognition and intent models on historic calls of the operator, CRM and billing context before every call, live agent assistance with transcripts and intent, analytics from both successful and failed calls.