Video course · 11 chapters · 23 min · certificate
AI Copilot on Snowflake Cortex AI: Customer 360 & Call Analytics
How an AI agent copilot and call-quality analytics app was built on Snowflake Cortex AI for insurance sales: the cross-selling problem, data architecture, Customer 360, life-event detection, recommendations, scripts, and call analytics for coaching.
What you'll learn
- Explain why cross-selling is hard for sales agents
- Describe a Customer 360 data architecture
- Use life events to trigger recommendations
- Generate ranked product recommendations and scripts
- Analyse call quality with transcripts and sentiment
- Use analytics to coach agents
Chapters
11 chapters · 23:09-
2:11
01Intro Members
Overview and business value
An AI agent copilot and call center quality analytics app built on Snowflake.
-
2:50
02Problem Members
Why cross-selling is hard
Information overload and other factors make cross-selling difficult for agents.
-
1:32
03Architecture Members
Data architecture
Ingesting raw customer data into a unified platform.
-
2:56
04Capabilities Members
Customer 360 capabilities
A Customer 360 view drives AI product recommendations with ranking.
-
2:34
05Life events Members
Tech stack and life events
The platform stack and detecting life events that trigger needs.
-
1:27
06Engine Members
The recommendation engine
Adding parameters like premium, discounts, claim history and loyalty.
-
1:52
07Scripts Members
AI script generation
Generating ranked products with propensity scores and talking scripts.
-
1:06
08Interface Members
The agent interface
A Streamlit app with agent info, customer search, profile and insights dashboard.
-
2:35
09Demo Members
Generating recommendations
A detected move to a new home triggers new recommendations.
-
1:39
10Analytics Members
Call quality analytics
Call metrics, sentiment, transcripts, translation and extracted entities.
-
2:27
11Coaching Members
Coaching and conclusion
Using analytics to coach agents and improve, and closing thoughts.
Jobs this course suits
Our AI checked this course against 500 jobs; these get the most out of it. Each job links to its learning path.