Why Study with PlanetCert?
The Latest Questions
Practice questions and exam topics aligned with the current exam objectives.
Detailed Explanations
Go beyond the answer. Master the material with comprehensive learning and professional explanations for every concept.
AI-Powered Insights
Personalized preparation guidance that adapts to your performance and identifies weak spots automatically.
Exam Information
Official specifications published by Snowflake
Exam Format
Registration
Skills this practice test covers
- Explain what Snowflake Cortex AI offers and where it fits a business use case.
- Call AI functions such as AI_COMPLETE, AI_CLASSIFY and AI_FILTER from SQL.
- Ground model answers in company documents with a Cortex Search service.
- Control access to Cortex features with roles and keep sensitive data out of prompts.
- Extract text and fields from PDFs with AI_PARSE_DOCUMENT and Document AI.
Explore related topics
GES-P02 exam content
4 domains · 11 topics · 150 questions
Snowflake for Gen AI Overview18%26 questions
This domain checks the big picture: what Snowflake Cortex AI offers, which use cases it suits, and why it matters that models run next to governed data. It also touches newer features such as Snowflake Intelligence and the hosting of open-source models.
Skills tested
- Describe the main Snowflake Cortex AI features and the business problems each one solves.
- Explain why running models inside Snowflake keeps data under its existing governance.
- Recognize newer capabilities such as Snowflake Intelligence, Cortex Code and Model Context Protocol (MCP) support.
- Host and version an open-source model with Snowpark Container Services and the Snowflake Model Registry.
Questions by topic
- Snowflake Cortex AI overview and the Data Cloud approach to Gen AI10
- Gen AI use cases on Snowflake8
- Snowpark Container Services and Snowflake Model Registry7
- Prompt engineering, fine-tuning, and Model Context Protocol (MCP)1
Snowflake Gen AI Functions38%57 questions
The largest domain is about using AI from SQL and building on it. Candidates match an AI function to a task, ground answers with Cortex Search, answer questions over tables with Cortex Analyst, and improve output with prompt design or fine-tuning.
Skills tested
- Choose between AI_COMPLETE and task functions such as SUMMARIZE, TRANSLATE and SENTIMENT.
- Filter and label rows with AI_FILTER and AI_CLASSIFY inside a query.
- Build retrieval-augmented generation on a Cortex Search service.
- Give Cortex Analyst a semantic model so that it can turn questions into SQL.
- Decide when better prompts are enough and when Cortex Fine-Tuning is worth the cost.
Questions by topic
- Task-specific and general LLM functions16
- AISQL and the expanded AI function suite15
- Cortex Search and Cortex Analyst14
- Prompt engineering, fine-tuning, and Model Context Protocol (MCP)12
Snowflake Gen AI Governance29%45 questions
This domain covers who may use Gen AI features and how that use stays safe and affordable. Candidates grant access through roles, protect sensitive data, turn on Cortex Guard and track credit use.
Skills tested
- Grant Cortex access through the CORTEX_USER role and role-based access control.
- Keep personal data out of prompts and answers with masking and AI_REDACT.
- Turn on Cortex Guard to filter unsafe model responses.
- Decide when to allow cross-region inference for a model that the account's region lacks.
- Monitor Gen AI credit use in the ACCOUNT_USAGE views.
Questions by topic
- RBAC, privileges, and safe data handling18
- Cortex Guard, cross-region inference, and usage monitoring27
Snowflake Document Processing15%22 questions
This domain turns documents into data. Candidates build pipelines that read files from a stage, extract text and layout, and pull structured fields from PDFs and images.
Skills tested
- Extract text and layout from staged documents with AI_PARSE_DOCUMENT.
- Pull named fields such as dates and totals from forms with Document AI.
- Build, manage and tune a document parsing pipeline for cost and accuracy.
Questions by topic
- Document AI and document parsing functions22
Requirements
- No formal prerequisites: anyone can take this practice exam.
- Recommended: at least one year of building Gen AI solutions on Snowflake in an enterprise.
- Recommended: SQL and data engineering knowledge, and ideally some Python.
About the GES-P02 exam
GES-P02 is the practice exam that Snowflake itself offers for the SnowPro Specialty: Gen AI certification. It is built for the version of the certification exam that Snowflake released in May 2026. Snowflake builds its practice exams to the same specifications and domain weightings as the live certification exams. The four domains cover Snowflake's Gen AI capabilities, the Cortex AI functions, governance and document processing. Passing GES-P02 earns no credential; the certification comes only from the live exam.
The PlanetCert practice test covers the same four domains and gives the most questions to Snowflake Gen AI Functions, the largest domain. The explanations are most useful where two functions look alike, such as AI_COMPLETE for open prompts and AI_CLASSIFY for sorting text into set categories. Study Mode suits the governance questions, where roles, Cortex Guard and cross-region inference need a careful read. Brainy AI Assistant can explain when Cortex Search or Cortex Analyst is the better fit for a question about company data.
How to earn the certification
- Earns
- No certification (GES-P02 is a practice exam)Specialty practice exam
How to prepare for GES-P02
- 1
Start from Snowflake's study guide
Snowflake publishes a study guide for the Gen AI certification. Read it next to the four domain names and list every Cortex feature that it mentions.
- 2
Run each AI function in a lab
Call AI_COMPLETE, AI_CLASSIFY, AI_FILTER and older task functions such as SUMMARIZE on a small table. Note what each one returns and which privileges each call needs.
- 3
Build a search and an analyst service
Snowflake Gen AI Functions is the largest domain. Create a Cortex Search service over a few documents and a Cortex Analyst semantic model over one table, then compare the answers each one gives.
- 4
Rehearse the governance settings
Grant and revoke the CORTEX_USER role, turn on Cortex Guard, and check the cross-region inference setting. Then find the credit use of each call in the ACCOUNT_USAGE views.
- 5
Parse documents, then take a timed run
Load a few PDFs to a stage and extract them with AI_PARSE_DOCUMENT. Then take the GES-P02 practice test in Timed Mode, and ask Brainy AI Assistant about any function whose output was a surprise.
Start a timed GES-P02run
Snowflake recommends its instructor-led Snowflake Gen AI Training course for this certification.
Who GES-P02 is for
- Engineers who build chatbots, search or document pipelines with Snowflake Cortex AI.
- Snowflake users who are preparing for the SnowPro Specialty: Gen AI certification.
- Data platform teams that must keep AI access, sensitive data and credit costs under control.
Typical job titles: Data engineer · AI engineer · Analytics engineer · Snowflake developer · Data architect
GES-P02 glossary
19 terms from the exam objectives and the vendor's official material, each defined in a sentence.
- AI_CLASSIFY
- An AI function that sorts text or images into categories that the user defines.
- AI_COMPLETE
- The general AI function that sends a prompt to a chosen large language model and returns the generated text.
- AI_FILTER
- An AI function that returns true or false for a plain-language condition, so that it can filter rows in a query.
- AI_PARSE_DOCUMENT
- An AI function that reads a document file and returns its text, with an option to keep layout such as tables.
- AI_REDACT
- An AI function that finds personal information in text and masks it.
- AI_TRANSCRIBE
- An AI function that turns the speech in audio files into text.
- Cortex Analyst
- A Snowflake service that turns plain-language questions into SQL over structured data, guided by a semantic model.
- Cortex Guard
- An option that filters harmful or unsafe content out of large language model responses.
- Cortex Search
- A Snowflake service for keyword and vector search over text, often used to ground model answers in company documents.
- CORTEX_USER role
- The Snowflake database role that allows calls to Cortex AI functions; administrators grant it to other roles.
- Cross-region inference
- A setting that lets Snowflake process an AI request in another region when the account's own region lacks the model.
- Document AI
- A Snowflake feature that extracts structured fields, such as dates and totals, from documents like invoices and forms.
- Model Context Protocol (MCP)
- An open protocol that lets AI agents connect to outside tools and data sources in a standard way.
- Retrieval-augmented generation (RAG)
- A method that first finds relevant passages and then passes them to the model, so that its answer rests on real documents.
- Semantic model
- A description of tables, columns and business terms that tells Cortex Analyst how to write correct SQL.
- Snowflake Cortex AI
- Snowflake's built-in set of AI features, including model functions and search and analysis services, that run inside the platform.
- Snowflake Intelligence
- Snowflake's agent experience that lets business users ask questions in plain language and get answers from their data.
- Snowflake Model Registry
- A store inside Snowflake for logging, versioning and serving machine learning models.
- Snowpark Container Services
- A Snowflake service that runs containers, such as an open-source model, on compute pools inside the account.