Every query in this book has been something you could paste into a SQL console and run once. A real experimentation pipeline needs those same queries to run on a schedule, in order, with the incremental logic from Chapter 7 handled automatically and a failing SRM check actually stopping bad data from reaching a dashboard. dbt is the standard tool for turning a folder of queries into that pipeline.
Why Not Just Schedule the Raw SQL
Nothing in this book requires dbt specifically — every query so far runs as plain SQL...