How Checkr cut an hours-long metrics investigation to minutes by engineering AI context on dbt and Omni

Checkr described how it improved AI-driven analytics by adding structured business context to its dbt and Snowflake environment, where more than 6,000 models had previously led agents to produce inconsistent and incorrect metric answers. Its approach distinguishes among the data warehouse, a semantic layer for definitions and valid calculations, and an AI context layer containing synonyms, sample values, descriptions, behavioral guidance, and example queries. Checkr automated context improvement through evaluation sets based on real stakeholder questions, AI-assisted failure clustering and context retrieval, retesting, and subject-matter-expert review, while also outlining a simpler manual workflow using existing documentation and dbt code. In one example, this context enabled an agent to identify international queues as the cause of First Reply Time SLA underperformance in minutes rather than hours. The session also cautioned that adding context indiscriminately can increase response time and token costs without improving accuracy, recommending targeted additions tied to specific, measurable failures.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.