Building Powerful AI Data Tools with AI Functions

Most enterprise data lives in contracts, invoices, call recordings and images. That data is hard to query and expensive to process with AI at scale. In this hands-on lab, you'll turn that unstructured content into structured, analysis-ready data using only SQL. Then you'll tune your pipeline for accuracy and cost before taking it to production.

During this instructor-led session, you will:
 

• Process documents at scale. Parse and extract fields from PDFs such as invoices or contracts with AI_PARSE_DOCUMENT and AI_EXTRACT. Then route and categorize them with AI_CLASSIFY.
• Analyze audio and images. Transcribe call recordings with AI_TRANSCRIBE, analyze images with multimodal AI_COMPLETE, and summarize themes across thousands of records with AI_SUMMARIZE_AGG.
• Evaluate and optimize with AI Function Studio. Measure output quality against a labeled dataset, tune prompts and models for accuracy, and estimate and control spend before you scale.

You'll leave with a working end-to-end pipeline, a reusable evaluation workflow and a clear way to choose between quality and cost for your own use cases. Familiarity with SQL is recommended. No ML experience is required.