| name | topdesk-ai-governance-cockpit |
| description | Design Power BI governance cockpits for TOPdesk AI/KI features. Use for AI monitoring datasets, semantic models, DAX measures, report pages, prompt/model version tracking, suggestion acceptance, override reasons, confidence drift, cost, token usage, SLA impact, PII/policy findings, feedback loops, evaluation results, RLS, and audit reporting. |
TOPdesk AI Governance Cockpit
Use this skill when AI/KI usage needs measurable governance, operational monitoring, and executive reporting.
Workflow
- Define monitoring grain: suggestion, prompt run, feedback event, evaluation run, source chunk, or cost event.
- Model facts and dimensions for Power BI: AI suggestions, prompt runs, feedback, evaluations, embeddings/source chunks, costs, users, dates, branches, feature type.
- Add DAX measures for adoption, quality, risk, cost, and business impact.
- Design pages: executive governance, operator adoption, quality/evals, cost, risk/compliance, feature drill-through.
- Apply RLS and PII minimization before exposing examples or prompts.
- Define refresh, retention, anomaly alerts, and release gates.
References
- Load
references/governance-cockpit-model.md for model and report architecture.
- Load
references/governance-kpis.md for KPI definitions and DAX guidance.
Assets
- Use
assets/governance-cockpit-spec-template.md.
- Use
assets/ai-governance-measures.dax.
- Use
assets/ai-governance-theme.json.
Scripts
- Use
scripts/new_governance_cockpit_spec.py to generate a cockpit spec.
