Persona: Analyst (Explore, Validate, Explain)

This page is a practical workflow for analysts using MCP Engine to understand semantic models, validate numbers, and explain “why is this KPI doing that?”.

What you typically want

  • Quickly understand model structure and key business measures.
  • Validate totals and breakouts with small DAX queries.
  • Compare security roles (when needed) without exposing row-level data.
  • Generate documentation and query suites for stakeholders.
  1. Connect intentionally and confirm context

    "List available models/datasets and ask me which one to connect to (don't auto-select). Confirm Desktop vs Service." For the full connection workflow, see Connecting and Context.

  2. Get oriented (model tour)

    “Summarize the model: main tables, relationships, key measures, calc groups, and roles.”“Highlight likely fact tables and date tables.”
  3. Find what you care about

    “Search for ‘margin’ across names/descriptions; then show the top 10 relevant measures with short explanations.”“If needed, search expressions and list the measures/calc items that implement the logic.”
  4. Validate numbers (small, safe queries)

    “Run a validation query for [Total Sales] by month for the last 12 months. Return only aggregates.”“Return top 20 products by [Gross Margin %] for last 30 days. Keep it small.”
  5. Explain results (with evidence)

    “Explain [Gross Margin %] and run a tiny query that shows numerator/denominator side-by-side for top 10 products.”

Prompt patterns that work well for analysts

Common analyst use cases (copy/paste)

“What measure drives this visual?”

“Find the measure behind ‘Gross Margin’ (table + name). Explain it in plain English and list dependencies.”

“Are the totals correct?”

“Validate [Total Sales] last 30 days vs prior 30 days. Return only aggregates and explain any unusual deltas.”

“Is security causing differences?”

“Test this aggregate query across roles A/B/C and summarize differences (no raw rows).”

Locked-down environments

If you're in browse-only or read-only mode, see Modes and restrictions for what's available and fallback strategies.

Risk & governance reminders

See also