Nuance contract expires in 9 months

Signed by week 14 → first markets live inside the window

Pre-LLM dialog engine — multi-turn AI impossible

Voice-native STT + TTS + Voice Agents, sub-300ms dialog

Cost ceiling blocks 100% QA / compliance coverage

60–80% lower per-minute → 100% coverage in the same budget

Genesys routing + Salesforce CRM must remain

We integrate into both — no rip-and-replace

Non-English failures — PT, Mandarin, Hindi

Best-in-class WER in those languages; custom-trained in days

Per-minute ceiling prevents voice AI on all calls

Ceiling removed — every contained call ~$5.25 cheaper

Latency + architecture block agents, GenAI, assist

Sub-300ms unlocks all three — one platform, Y1–Y3

RFP: Deepgram vs. AssemblyAI vs. self-hosted Whisper

The only full voice-agent platform in this RFP

Year 1: routing + self-service in 6–8 markets

Phased, gated rollout — every market validated before cutover

Y2–3: agent assist, voice analytics, GenAI flows

Same platform — pre-priced at signature, zero re-procurement

AWS eight-figure EDP · SOC 2 + VPC required

Private Offer = budget-neutral, 6–10 wks faster · SOC 2 II