| name | mcp-engine-ai-readiness |
| description | Use when preparing or assessing a Power BI semantic model for Copilot, Fabric data agents, or natural-language Q&A — clear business terminology, unambiguous metrics, usable date defaults, focused field exposure, descriptions, AI instructions, AI data schema recommendations, verified-answer candidates, or natural-language validation tests. For general modeling quality unrelated to AI consumption, use mcp-engine-model-quality. |
PBI AI Readiness
Use this skill to turn an existing Power BI semantic model into a Copilot-ready assessment and artifact pack. This is an authoring and review workflow, not a runtime MCP tool.
Start Here
- Confirm the user wants a readiness assessment, artifact drafts, or both.
- Prefer metadata-level inspection before querying data values.
- Use existing SemanticOps MCP tools when available:
list_model, manage_dependencies, run_query, manage_tests, and manage_model_properties.
- Keep unsupported Prep data for AI actions as drafts for Power BI Desktop, Power BI service, PBIP, Git, or manual review.
- Separate recommendations into:
- can apply through MCP/model metadata now
- draft/export for Prep data for AI UI or PBIP/Git workflow
- validate manually in Copilot
Workflow
- Read copilot-readiness-workflow for the assessment sequence, tool usage, privacy guardrails, and output order.
- Read readiness-scorecard when producing severity, score, business impact, and remediation priority.
- Read ai-artifact-templates when drafting AI instructions, AI data schema recommendations, verified answers, or
manage_tests candidates.
- Read domain-examples when the model is sales, finance, support, or operational and the user wants concrete starting examples.
Guardrails
- Do not claim SemanticOps MCP can directly configure all Power BI Prep data for AI settings over live TOM/XMLA.
- Treat AI instructions, AI data schemas, and verified answers as draft artifacts unless the user provides an explicit supported PBIP/Git path or asks for manual-application guidance.
- Do not expose sensitive values from data previews. Prefer names, descriptions, expressions, relationships, dependencies, and aggregate-only validation queries.
- Respect SemanticOps MCP mode, policy, confirmation, license, and audit gates for any suggested or requested model change.
- Make nondeterminism explicit: readiness work can improve Copilot behavior, but it cannot guarantee identical answers for every prompt.
Output Standard
Return a compact readiness pack unless the user asks for raw details:
- Executive summary with readiness level.
- Scorecard grouped by critical, high, medium, and low findings.
- Recommended MCP-applicable model metadata fixes.
- Draft AI instructions.
- Draft AI data schema recommendation.
- Verified-answer backlog.
- Optional natural-language test suggestions.
- Manual validation checklist for Power BI Desktop or service.
