AI Sales Agent
Problem: Customers needed fast, accurate answers on courses/licensing/policies from a huge, changing catalog; support handled repetitive questions and compliance is state-specific.
Solution: Built RUBI, an embedded RAG purchase advisor in using ETL + multimodal indexing and hybrid retrieval (metadata + vector + structured lookups) with cited, grounded answers, purchase flows, analytics, and tuning loops.
Metrics: 695,042 conversations; 2.6 Qs/convo; feedback 6% up / 94% down; ~10s to first token, ~11s complete; KB >100k structured rows.
0 likes · 3 views · Jan 5, 2026
AI Sales Agent Problem: Customers needed fast, accurate answers on courses/licensing/policies from a huge, changing catalog; support handled repetitive quest...
