CCA-F: Claude Certified Architect — Foundations

You are developing a structured data-extraction system with Claude. The system pulls information from unstructured documents, validates its output with JavaScript Object Notation (JSON) schemas, and preserves high accuracy. It must gracefully handle edge cases and integrate with downstream systems.

Your system has operated for 3 weeks, and human reviewers have corrected 847 extractions. Analysis identifies a recurring pattern: when recipes contain informal measurements such as “a handful” or “a splash,” the model either fabricates specific quantities or leaves fields blank—representing 23% of all corrections.

How should this feedback be used to improve extraction accuracy?