ABAmine Bouhlal

// Case study · 2024 — present

Voice AI at logistics scale

A European parcel-logistics leader

Delivered as Solutions Architect at NTT DATA

180K+

calls / month in production

24/7

availability, no queue scaling

E2E

owned from discovery to deployment

CALLERTELEPHONYSIP ingressVOICE AGENTSTT · LLM · TTSELEVENLABSCARRIER APIstracking · ordersHUMAN AGENTcontext attachedRESOLVED
Call flow: telephony ingress → speech understanding → LLM dialogue grounded in carrier APIs → resolution or context-rich human handoff.

The problem

A contact centre absorbing six-figure monthly call volumes for repetitive, fully structured requests — parcel status, redelivery, address changes — with cost, wait times and agent churn scaling linearly with volume.

The system

Designed an end-to-end voice agent on ElevenLabs conversational AI: telephony ingress, real-time speech understanding, LLM-driven dialogue grounded in the carrier's tracking and order APIs, and a clean escalation path that hands complex or sensitive cases to human agents with full context attached. Owned the architecture from discovery through production deployment.

The outcome

Live in production handling 180,000+ calls per month, resolving routine requests without human involvement and shrinking the queue that reaches the human team to the cases that actually need judgment.

ElevenLabsPythonGCPVertex AIFirebase
Next case → Conversational AI for consumer financeDiscuss a similar system