Overview

The Power BI integration enables Prizm to connect to the Power BI Service (including Microsoft Fabric workspaces) as a downstream consumer of your data. Like Tableau, Power BI is a BI/pipeline integration — Prizm ingests semantic model, report, and dashboard metadata to extend lineage from warehouse tables through to the assets your business users query, and propagates quality scores, alerts, and issues downstream to the reports that depend on your data.

What Prizm Connects To

Prizm connects to workspaces backed by a capacity tier that exposes the XMLA endpoint:

Capacity Tier	Details
Premium Per User (PPU)	Per-user Premium license; XMLA endpoint available on assigned workspaces
Premium (Per Capacity)	Organization-wide Premium capacity; XMLA endpoint available on all workspaces assigned to the capacity
Microsoft Fabric Capacity	Fabric-licensed capacity; XMLA endpoint available on all workspaces assigned to the capacity

A standard Power BI Pro license alone does not expose the XMLA endpoint. At least one of the capacity tiers above must be assigned to every workspace Prizm will scan for semantic model metadata. See Setup for the full prerequisites.

Supported Power BI Objects

Prizm ingests the following object types from Power BI. The hierarchy reflects how objects are organized in the Power BI Service:

Workspace
 ├── Dataflow (optional, when Extract Dataflows is enabled)
 ├── Semantic Model
 │    ├── Table
 │    │    └── Column
 │    ├── Measure
 │    └── Relationship
 ├── Report
 └── Dashboard

Object	What Prizm Captures
Workspace	Name, type, state, linked capacity
Semantic Model	Name, owner, storage mode, created/modified timestamps, refresh history
Table / Column	Schema, data type, hidden state, lineage tag, upstream M expression
Measure	DAX expression, data type, format string, description
Relationship	Source/target table and column, cardinality, cross-filter direction
Report	Name, owner, URL, source semantic model
Dashboard	Name, owner, URL, tiles and their source reports/semantic models
Dataflow	Name, owner, entities, M expressions (when Extract Dataflows is enabled)

Observability Metrics

Prizm computes the following out-of-the-box metrics for Power BI assets:

Operational
Usage

Metric	Computed At	Description
Freshness	Semantic Model	Time elapsed since the last successful refresh for the semantic model
Reliability	Semantic Model	% of the semantic model’s refreshes that completed successfully
Job Duration	Semantic Model	Average refresh time across the semantic model’s successful refreshes

Freshness, Reliability, and Job Duration are computed once at the Semantic Model level. The same value is then propagated as-is to every Report built on that model — reports do not have their own independently computed values for these metrics.

Metric	Level	Description
Report View Count	Report	Number of times the report was opened, over a 14-day or 28-day window
Report View Count (Cumulative)	Semantic Model	Sum of the Report View Count across every report built on the semantic model, for the same window

Usage is tracked only at the Report level (plus the semantic-model-level rollup above) — Prizm does not compute a separate usage metric for Dashboards or Semantic Models directly. There is no 90-day window, and Prizm does not track Unique Users or Dataset Queries.

Metric defaults by asset type

Metric	Semantic Model	Report	Dashboard
Freshness	✅ Computed	✅ Inherited from semantic model	—
Reliability	✅ Computed	✅ Inherited from semantic model	—
Job Duration	✅ Computed	✅ Inherited from semantic model	—
Usage (Report View Count)	✅ Cumulative rollup	✅ Computed (14-day / 28-day window)	—
Column-level Profile / Structural	✅	—	—

Quality Propagation

Prizm can propagate data quality signals from upstream warehouse tables to the Power BI semantic models, reports, and dashboards that consume them. This is configured per connector at setup time.

Signal	Propagation Options
Score	Based on associated upstream table, or none
Alerts	Based on associated upstream table, or none
Issues	Based on associated upstream table, or none

When propagation is enabled, a quality issue on a Snowflake or Databricks table automatically surfaces on every semantic model, report, and dashboard that draws from that table — giving report consumers visibility into upstream data health without leaving Power BI.

Criticality

Power BI assets are scored using Prizm’s Criticality framework, applied with the following Power BI-specific formulas. Each input is log-normalized (ln_norm) — scaled to a 0–1 score over the stated range — before being weighted.

Report

report_raw_score =
    0.50 × ln_norm(views, 1..20)
  + 0.50 × ln_norm(dashboard_count, 0..10)

views — the report’s Report View Count (see Observability Metrics above)
dashboard_count — number of dashboards that include this report

Dashboard

A dashboard’s criticality is inherited from its most critical report — the highest report_raw_score among the reports it includes.

Semantic Model

sm_raw_score =
    0.70 × ln_norm(highest_report_tier, 0..4)
  + 0.15 × ln_norm(report_count, 0..20)
  + 0.15 × refresh_reliability

highest_report_tier — the criticality tier (0–4) of the semantic model’s highest-scoring report
report_count — number of reports built on this semantic model
refresh_reliability — the semantic model’s Reliability metric (see Observability Metrics above)

Criticality determines which semantic model columns the Profile Recommendation and Structural Recommendation jobs select for column-level monitoring — higher-criticality columns are profiled more frequently.

Supported Features

Feature	Supported
Service Principal authentication	✅ Recommended
Master User authentication	✅
OAuth 2.0 authentication	✅
Semantic model schema via XMLA	✅
Report / Dashboard ingestion	✅
Dataflow ingestion	✅ (optional)
Freshness / Reliability monitoring	✅
Usage analytics	✅
Column-level profiling	✅ (Premium/PPU/Fabric capacity required)
Quality score propagation	✅
Microsoft Fabric workspaces	✅
Authoring or modifying Power BI content	❌ Out of scope
Triggering semantic model refreshes	❌ Out of scope

Next Steps

Setup

Connect Prizm to Power BI — prerequisites, authentication methods, and step-by-step configuration.

What We Collect

Full field-level breakdown of every metadata object Prizm extracts from Power BI.

FAQ

Common questions about authentication, scoping, and permissions.

Glossary

Definitions for Prizm terms used throughout this guide.

Get Started

Core Concepts

Platform

Architecture

Data Sources

Integrations

Security & Access Control

Deployment

Help

Settings

MCP

Release Notes

API Reference

What Prizm Connects To

Supported Power BI Objects

Observability Metrics

Metric defaults by asset type

Quality Propagation

Criticality

Report

Dashboard

Semantic Model

Supported Features

Next Steps

Setup

What We Collect

FAQ

Glossary

​What Prizm Connects To

​Supported Power BI Objects

​Observability Metrics

​Metric defaults by asset type

​Quality Propagation

​Criticality

​Report

​Dashboard

​Semantic Model

​Supported Features

​Next Steps

Setup

What We Collect

FAQ

Glossary

What Prizm Connects To

Supported Power BI Objects

Observability Metrics

Metric defaults by asset type

Quality Propagation

Criticality

Report

Dashboard

Semantic Model

Supported Features

Next Steps