Snowflake SnowPro Specialty: Gen AI (GES-C02)

Core principles, capabilities, and best practices of Snowflake's generative AI stack: where Cortex runs, how data stays in the account's governance boundary, model selection, and typical Gen AI workflows and architectures.

• Snowflake Cortex as fully managed, in-account LLM inference (no data leaves Snowflake)
• Cortex AI building blocks: LLM functions, Cortex Search, Cortex Analyst, Cortex Agents
• Serverless vs. warehouse compute for Cortex; credit-based consumption
• Choosing a model by task, quality, latency, and cost; available model families
• RAG pattern on Snowflake: chunk, embed, retrieve, ground the prompt
• SQL and Python (Snowpark) entry points to Cortex
• Cortex capability map: which problem each service solves (Search = retrieve from text, Analyst = query structured data, Agents = reason and orchestrate)
• Snowflake CoCo (formerly Cortex Code) — AI coding agent in Snowsight and the CLI, for SQL/Python development and administration
• Snowflake CoWork (formerly Snowflake Intelligence) — conversational work agent over governed enterprise data for non-SQL business users
• CoCo vs Cortex Agents: an assistant for building on Snowflake vs an application that uses tools to execute business tasks
• Ways to interact with Cortex: AI Studio, SQL, and REST APIs
• Managed Cortex capabilities vs bringing your own model (Model Registry, Snowpark Container Services)

Free Snowflake for Gen AI Overview practice questions →Snowflake for Gen AI Overview study guide →

Hands-on use of Cortex LLM and AISQL functions, embeddings and vector search, Cortex Search for retrieval, Cortex Analyst for text-to-SQL, and fine-tuning — the deepest and most heavily tested area.

• COMPLETE / AI_COMPLETE for generation, prompting, and structured (JSON) output
• Task-specific functions: SUMMARIZE, TRANSLATE, SENTIMENT, EXTRACT_ANSWER, CLASSIFY_TEXT / AI_CLASSIFY, AI_FILTER, AI_AGG
• EMBED_TEXT_768 / EMBED_TEXT_1024, the VECTOR type, and VECTOR_COSINE_SIMILARITY
• Cortex Search service for semantic/hybrid retrieval powering RAG
• Cortex Analyst and semantic models for natural-language querying (text-to-SQL)
• Cortex Fine-tuning (FINETUNE) to adapt base models to a task
• Cortex AI Functions (formerly Cortex AISQL) and the AI_* family: AI_COMPLETE, AI_CLASSIFY, AI_EXTRACT, AI_FILTER, AI_SENTIMENT, AI_AGG
• Cortex Analyst semantic views: dimensions, facts and metrics, and why a semantic layer is required for text-to-SQL
• Verified queries and Semantic View Autopilot for improving Analyst accuracy
• Cortex Agents: available tools, combining Analyst and Search in one request, and when an agent beats a single capability

Free Snowflake Cortex AI and LLM Functions practice questions →Snowflake Cortex AI and LLM Functions study guide →

Securing and governing Gen AI workloads: role-based access to Cortex, controlling which models can run and where inference happens, safety filtering, and applying Snowflake's existing data governance to AI pipelines.

• SNOWFLAKE.CORTEX_USER database role and granting access to Cortex functions
• Cross-region inference (CORTEX_ENABLED_CROSS_REGION) and data residency implications
• Restricting available models with the CORTEX_MODELS_ALLOWLIST parameter
• Cortex Guard for filtering unsafe or harmful model responses
• RBAC, masking policies, and row access policies on source and embedding tables
• Cost governance: resource monitors, budgets, and monitoring Cortex usage

Free Snowflake Gen AI Governance practice questions →Snowflake Gen AI Governance study guide →

Extracting text and structured data from documents to feed Gen AI pipelines using Snowflake's Cortex document-processing functions (the successor to the retired Document AI UI experience).

• PARSE_DOCUMENT / AI_PARSE_DOCUMENT for OCR and layout-aware text extraction
• OCR vs. LAYOUT modes and when to use each
• Reading source files from internal/external stages and directory tables
• Chunking extracted text (SPLIT_TEXT_RECURSIVE_CHARACTER) for embedding
• Building a document-to-RAG pipeline: parse, chunk, embed, index, retrieve
• AI_EXTRACT / structured field extraction from parsed documents

Free Document Processing practice questions →Document Processing study guide →