Chevron Blue Right Back to all events
working-with-data

Workshop: From terminal to dashboard with Metabase MCP + CLI

· 45 minutes

 

Guests

Stephen Tracy

Stephen Tracy

CDO & Udemy Course Creator

Stephen is a data and analytics practitioner with 15+ years of experience across data science, market research, and AI. He's a former Chief Data Officer at North One and co-founder of Milieu Insight, and has taught tens of thousands of students through his data courses. He also hosts the Empirical Storytellers podcast, where he talks with data and AI leaders about how decisions actually get made.

Pierina Salinas

Pierina Salinas

Product Advocate, Metabase

Summary

In this workshop, Stephen Tracy and Pierina Salinas explored how AI is changing data work and when to use MCP vs. CLI workflows with Metabase.

What’s changing in data work

  • Roles are blending — data engineers, analysts, and scientists increasingly wear multiple hats.
  • Technical skills are commoditized — more people can generate SQL and dashboards with AI prompts.
  • Judgment is now the bottleneck — reviewing and iterating on AI output matters more than raw execution speed.
  • Self-service analytics is becoming practical for non-technical stakeholders.
  • Generative BI introduces governance, quality, and security tradeoffs that teams need to manage.

MCP vs. CLI: when to use each

MCP (Model Context Protocol)

Best for:

  • Exploratory analysis and Q&A
  • Interactive charts and dashboards
  • Non-technical users
  • Quick insights

Not ideal for:

  • Complex multi-step workflows
  • Database migrations
  • Batch operations

CLI (Command Line Interface)

Best for:

  • Warehouse migrations
  • ETL transforms
  • Bulk operations
  • Infrastructure tasks

Not ideal for:

  • Quick exploratory queries
  • Non-technical users
  • Interactive visualization

Key insights

  • “Execution is table-stakes” — hard technical skills are expected, not differentiating.
  • “Judgment > Speed” — the ability to validate AI output is now a core differentiator.
  • “Semantic layers are insurance” — upfront modeling work reduces cost and downstream errors.
  • “MCP for humans, CLI for machines” — choose tools based on audience and workflow.
  • “Reference dashboards = 40-60% better output” — examples improve AI-generated results.