You'll start by meeting CoCo and getting around a session, then use it to explore a retailer's database, check its answers with your own queries, and control what it can change with permission modes and plan mode.

Meet CoCo

Why teams pick the CLI

Getting around a session

Fixing an inherited query

Exploring an unfamiliar database

Let CoCo map the data

Ask with what you know

Check the answer yourself

Permissions and plan mode

Approve CoCo's first change

Plan before you build

Look without touching

Your AI Assistant in the Terminal

Next, you'll have CoCo run Snowflake's AI functions over customer reviews, turn your work into scripts that run without a session, and build pipelines with dynamic tables and scheduled tasks that keep running on their own.

AI workflows in SQL

Find what's behind low ratings

Run a skill on a new question

Why the team uses skills

Scripting with CoCo

Script the refund check

Why keep the script

Ask CoCo from the shell

CoCo in pipelines

Keep review topics up to date

Why the weekly summary didn't run

Pick the right pipeline

Congratulations!

Bonus: Real-time fraud detection with CoCo

Automating AI Workflows with CoCo

CoCo is Snowflake's AI coding agent, and its CLI brings it into your terminal, connected to your Snowflake account. This course is hands-on: most exercises put you in a real terminal with CoCo, working on live data in Snowflake. You'll explore an unfamiliar retail database, check the answers CoCo gives you, and decide what it's allowed to change. Then you'll have it run AI functions over customer reviews, write scripts you can run again, and build pipelines that keep working after you close the terminal. By the end, you'll know how to hand CoCo real data work and stay in control of what it does.

Introduction to Snowflake

Put Snowflake CoCo to work in your terminal: explore data, run AI functions over text, and build scripts and pipelines while you stay in control.

Introduction to Snowflake CoCo CLI

Why teams pick the CLI

Exercice interactif pratique