search
Home

Best Practices

hashtag
Best Practices to split the scope by Environment

  1. The Medallion Approach: In Connecty AI, the Data Environment determines the scope for data quality, business rule learning, and query history. To ensure high-quality AI responses, we recommend splitting your scope into separate environments based on the target audience:

    1. The Business Consumption Environment (Gold Layer)

      1. Scope: Include only the finalized tables and views required for daily business dashboards, KPIs, and ad-hoc analysis.

      2. Goal: Dedicated to business teams. Ensuring the AI only accesses this layer prevents it from learning irrelevant patterns from raw data and ensures high-precision answers.

      3. Note: This corresponds to the Gold layer in the Medallion Architecture.

    2. The Engineering Environment (Bronze/Silver Layers)

      1. Scope: Include raw data tables and intermediate transformation views used for maintaining data pipelines.

      2. Goal: Dedicated to data engineering teams for lineage tracking and pipeline debugging. Keeping this separate ensures that technical data does not "pollute" the business context.

hashtag
Best Practices: Configuring Data Workspaces

After establishing your primary Data Environment, use Goals AI (automatically triggered during AI Sync process) to split your scope into domain-specific workspaces. This ensures the AI retrieves only the most relevant context for every query.

1. Leverage AI-Suggested Domains Instead of manually selecting tables, allow Connecty’s AI to analyze your schema and suggest Business Domains (e.g., Sales, Marketing, Supply Chain).

  • Action: Review the AI’s suggested domains. Deselect any domains that are not relevant to your current analytics needs.

2. Define Goals & KPIs Per Workspace Under each domain, the AI will suggest specific Business Goals and North Star Metrics.

  • Best Practice: Explicitly refine these goals (e.g., change "Track Sales" to "Optimize Recurring Revenue & Churn").

  • Impact: When you update a goal, the AI automatically regenerates the scope, selecting only the specific tables and columns needed to answer questions about that goal.

Last updated