02 · Applied AI + Self-Service Analytics
cabi Intelligence
Turning recurring ad-hoc data questions into governed, natural-language self-service analytics with Snowflake Cortex Analyst.
01
Semantic models
02
Dimensions
03
Metric logic
04
Validation
05
Business language
The problem
Analytics requests arrived from multiple teams, often requiring repeated interpretation of company-specific terminology, filters, dimensions, and metric definitions.
The real challenge
The hard part was not connecting an AI model to Snowflake. It was making the answers trustworthy. Semantic models, relationships, dimensions, filters, metric definitions, and business terminology all had to reflect how teams actually use the data.
What I built
I designed and validated semantic models, verified generated queries, established custom instructions, and incorporated company- and team-specific terminology so Cortex Analyst could return results aligned with business expectations across functions such as Marketing, Sales, and Field Operations.
Why it matters
The project turns the analytics team from a request queue into an enablement layer: business users can ask questions in natural language while the underlying system remains grounded in governed data and explicit business logic.
Confidentiality note
Company-specific implementation details and data are intentionally generalized for this public portfolio.
← Back to all work