Build Analytics Agents for Your Teams and Customers

What is "AI for BI" or "Conversational BI"?

Dashboards are designed to share key metrics and trends, but inevitably lead users to ask more questions. Conversational BI is putting a data analyst in the hands of every dashboard user to answer the more difficult, granular questions that the dashboard cannot answer.

This session is for Snowflake users who want to give their internal teams or their customers the ability to chat with their data to answer ad-hoc questions.

We walk through the core building blocks of building an agent (Semantic Views, Cortex Search, and Cortex Agents) and explain how each helps to provide accurate answers to the user. Then we build an end-to-end agent live and show it running in Snowflake CoWork (including Dashboards 2.0) and inside a customer-facing application.
In this Hands-on Lab, we will build an analytics agent step by step:
 

• Define a Semantic View so the agent understands the data

• Add Cortex Search so it can retrieve data from PDF documents

• Assemble a Cortex Agent in Cortex Code

• Test the agent in CoWork, Dashboards 2.0, and in a customer-facing app

Attendees will leave with a clear understanding of the Cortex AI architecture, how to build agents with Cortex Code or in the UI, and how to provide the agent to their teams or customers.