Snowflake Cortex AI: Cortex Code (CoCo), Search & Agents

This course teaches Snowflake Cortex end to end: the AI functions, the prompts, and the SQL behind each feature.

You'll work through Snowflake's Cortex platform feature by feature:

• Cortex Code (CoCo) — prompting Snowflake's AI coding agent in Snowsight to scaffold the pipeline the rest of the course runs on

• Cortex AI Functions (AISQL) — AI_COMPLETE, AI_CLASSIFY, AI_SENTIMENT, AI_SUMMARIZE_AGG, AI_TRANSLATE, AI_REDACT, and the prompts behind each one, for classifying, scoring, summarizing, translating, and redacting text directly in SQL

• Cortex Search — creating a search service, chunking strategy, and querying it, plus a working RAG pattern that combines Search with AI_COMPLETE

• Semantic Views and Cortex Analyst — modeling metrics and dimensions, then prompting Cortex Analyst with plain-English business questions and reading the SQL it generates

• Cortex Agents — writing agent instructions, combining Analyst and Search as tools, and working through the plan-and-reason loop

• Snowflake CoWork — the agent experience built for business users, and how the same governance rules carry over

• Prompting, RAG, and fine-tuning — a decision framework for choosing between them, plus a fine-tuning job run in Snowsight

Each feature is demonstrated hands-on against a small, consistent dataset (a car rental business), so the focus stays on the AI functions and prompt patterns themselves rather than on data modeling. Governance is covered alongside each feature: which role or privilege it needs, and what a locked-down user can and can't do with it.

In the final module, everything that has been built, will be brought together in one Streamlit Dashboard.

Note: This is not a Snowflake introductory course.