Can AI Face Your (Potential) Customers?
summary
Claims concrete wins: a custom speech‑to‑text pipeline trained with freelancers beat Azure and AWS in two weeks for Baltic languages, handles multi‑language mixes and dialects, detects language in ~4–5 seconds, and extracts 100+ data points to replace 15 minutes of post‑call work after a 5‑minute conversation. Architecture runs on‑premise with isolated networks, uses small specialized models for cheap scaling instead of big LLMs, and provides real‑time BI hooks and routing into powerbi, Genesis, ServiceNow while offering on‑premise, offline deploy options.
AI-written summary · about
chapters (10)
- 0:00AI in contact centers
- 1:04IBM Watson to East Lucid AI
- 3:31AI for sales and support
- 5:56Call center pain points
- 12:29AI solution architecture
- 16:30Speech-to-text for Baltic languages
- 24:23Call conversation insights
- 34:02Live AI form-filling demo
- 47:26GDPR and data protection
- 50:59Q&A
related talks
51mCan AI Face Your (Potential) Customers?
29mAI for All
46mEnhance your solutions with new Azure AI products and features
44mGive your GenAI apps a multilingual voice with Azure AI Speech
31mBuild AI-assisted communication workflows for customer engagement
49mHow to be an AI First company
56mAI Agents That Don’t Suck: How to Build Ones That Actually Drive Business Value
11mLeveraging AI and Microsoft Ecosystem for Talent Acquisition