Module 3 • Lesson 2440 mins

Define Indexing Cadence to Control Data Freshness

Trace a policy change's lifecycle through the RAG pipeline, compare 3 indexing cadence options, and write a freshness requirement around acceptable delay instead of a specific schedule.

Describe a source-data change's lifecycle until the RAG system reflects it
Compare real-time, batch, and manual indexing by risk level
Write a freshness requirement framed around acceptable delay, not a cron schedule

Define Indexing Cadence to Control Data Freshness

Lesson 19 assumed the chunks are already sitting in the system. But real-world source data (policies, documents, FAQs) changes over time — this lesson's question is: when the source changes, when does the RAG system find out, and how long does it keep answering from stale data in the meantime?