Snowflake Cortex, Explained Like an AI That Lives Next to Your Data Snowflake Cortex is a set of AI capabilities built directly into the Snowflake data warehouse, allowing users to run AI tasks like summarization, translation, and natural-language querying using simple SQL without data ever leaving the warehouse. The feature eliminates the need to move data to external AI services, reducing latency and security risks. Estimated reading time: ~11 minutes. No prior experience required. The first time I bolted AI onto a data warehouse, the plan looked reasonable on a whiteboard: pull the data out of the warehouse, send it to an AI service somewhere else, get the answer, bring it back . In practice it was clumsy. Data had to leave its secure home, travel across the network, get processed elsewhere, and return, slower, more expensive, and now I had sensitive data copied into a second place I had to worry about. The obvious question: why move the data to the AI at all, why not bring the AI to the data? That's exactly the idea behind By the end of this post you'll understand what Cortex is, its main capabilities, how you use it, the traps, and how it fits the modern data stack. Quick note: Snowflake and Cortex are commercial products, not open source. We include this post because "AI inside the warehouse" is a pattern you'll meet everywhere now, and theideashere apply to similar features in other platforms too. One sentence: Snowflake Cortex is a set of AI capabilities built directly into the Snowflake data warehouse, letting you run AI tasks, like summarizing text, translating, or asking questions in plain English, using simple SQL, without your data ever leaving Snowflake. If you've read the Snowflake post, this is "Snowflake, now with a brain attached." Imagine your company data is a huge, secure library, and you need documents analyzed. You could mail each document to an outside consultant, wait, and get it back, slow, and now your documents are sitting on someone else's desk. Or you could hire an in-house expert who works inside the library. Documents never leave the building. You just walk over and ask. Faster, safer, simpler. Cortex is that in-house expert. The AI lives where the data lives, so you ask questions right where everything already is, through the SQL you already use. Cortex bundles several AI abilities. Here are the ones you'll actually reach for. The simplest and most striking: you call an AI model like it's a normal SQL function. Want to summarize a column of customer reviews? It's one function call applied across your rows. Sentiment, translation, summarization, all as SQL you run right in the warehouse. -- Summarize each review, right inside a query conceptual SELECT review id, SNOWFLAKE.CORTEX.SUMMARIZE review text AS summary, SNOWFLAKE.CORTEX.SENTIMENT review text AS mood FROM customer reviews; No data export, no separate service, the AI runs where the rows already are. This is the "talk to your data" capability: a business user types a question in plain English, "what were our top five products by revenue last quarter?", and Cortex translates it into SQL, runs it, and returns the answer. It turns the warehouse into something a non-SQL user can interrogate directly. The key to making this reliable is a semantic model , a description of what your tables and columns mean in business terms, so the AI maps "revenue" to the right column instead of guessing. For question-answering over documents, Cortex Search finds the most relevant passages from your text data the retrieval piece of the RAG pattern from the LangChain post , built in, so you don't assemble a separate search system. The newest layer: agents that can combine the above, search for relevant data, generate SQL, call the LLM functions, to answer more complex, multi-step questions, all within the platform's security boundary. php flowchart LR U User asks in English -- A Cortex Analyst / Agent A -- SEM Semantic model: