At dbt Summit 2026, three analytics engineers from Kaizen Gaming - the company behind sports betting and casino brand Bettano, operating in 25 markets with over 15 million active users - presented the governance framework they built to keep dbt consistent across six analytics engineering domain teams.
As the organization grew, each domain team had adopted dbt at its own pace. The result was the same source data staged independently in as many as six places, shared infrastructure copy-pasted rather than reused, and no visibility for leadership into test coverage or investment payoff.
Rather than create a full platform team, the engineers centralized three layers: a Core project of company-wide spine models as lightweight views, an internal utilities package of shared macros, tests and configuration pushed automatically to subscriber repos, and infrastructure as code via Terraform with CI scoped to affected projects and merge rights limited to two code owners. They also built an enablement repository of roughly 16 analytics-engineering-specific skills wired into their warehouse and dbt's MCP server.
The reported results: new project lead time down 70%, $10,000 in annualized savings from templated project setup, roughly 15x time leverage on shared configuration fixes, and AI onboarding for new hires reduced from a multi-day process to about 30 minutes. Fifteen merge requests against shared repos in the 30 days before the talk were cited as evidence the framework is actively used.
The speakers offered a rule for what to centralize: pull something into the shared layer when no single domain owns it, when every team would otherwise solve the same problem independently, or when inconsistency is costly at the organization level. Leave it with the domain team when it encodes real business judgment or when errors affect only one team. On AI agent access, they said all analytics engineers currently share the same access level, with authentication for agents connecting through GitLab, Atlassian or Databricks MCP servers handled centrally via SSO. As one speaker put it, "governance and autonomy do not need to be in conflict."
Source: dbt Labs Blog