A post on dbt's official blog describes how modeling Gong call transcripts inside the data warehouse using dbt reduced AI-related token costs by 20x compared with pulling data directly through Gong's API1. The approach replaced repeated API calls to feed AI systems with pre-modeled warehouse tables, significantly lowering per-token expenditure.
dbt Blog: Warehouse-Based Transcript Modeling Cut AI Token Costs 20x vs. Gong API
A dbt blog post details how modeling Gong transcripts in the warehouse with dbt reduced AI token costs by 20x versus direct API access.