What problem does it solve? Adding LLM capabilities like classification, extraction, summarization, and translation to data pipelines normally requires managing model endpoints, API keys, and boilerplate code. This Skill shows how to call Databricks built-in AI Functions directly from SQL and PySpark so AI inference runs on table columns like any other function. ## Core Features & Use Cases - Task-Specific AI Functions: Use ai_classify, ai_extract, ai_summarize, ai_mask, ai_translate, ai_fix_grammar, ai_gen, ai_analyze_sentiment, and ai_similarity for batch inference without endpoint setup. - Document Processing & RAG: Chain ai_parse_document and ai_prep_search into staged Delta pipelines for parsing PDFs and building custom RAG indexes with Vector Search. - Advanced Querying & Forecasting: Use ai_query for custom endpoints, multimodal input, and structured JSON output, plus ai_forecast for table-valued time series forecasting. - Use Case: Parse a volume of vendor invoices with ai_parse_document, classify them with ai_classify, extract invoice number and total with ai_extract, and persist results to a Delta table for downstream analytics. ## Quick Start Use the databricks-ai-functions skill to classify the sentiment and extract key fields from the customer_feedback table using SQL.
