Get Started

Ask your Power BI estate anything. In plain English.

Understand every table, measure, and relationship, and generate full documentation automatically. AI works as your BI analyst, not your decision maker.

Book a Demo Call See Capabilities
🔒 Runs in your Azure tenant 🛡️ Connects to your existing workspaces ⚡ Documentation traced to source
"We have 60 reports and nobody who built half of them still works here. When a number looks wrong, it takes two days just to find which table it comes from, before anyone can even check if it's right."
10×
faster than manual
report documentation
100%
measures traced to
source tables
0
reports left
undocumented
The Solution

A BI analyst that never forgets how a report was built.

This isn't a chatbot bolted onto a dashboard, and it isn't a static data dictionary. It's a structured reasoning system with a full knowledge base over your tables, measures, visuals, and relationships, built by reading your actual Power BI models instead of a spreadsheet someone filled in by hand.

Ask it anything in plain English: why a number moved, what a measure means, which tables feed a visual, and it answers with the DAX, the lineage, and the business logic behind it. Point it at your workspace and it generates full documentation across hundreds of fields automatically, using your own naming conventions and templates.

Natural-language queries DAX & lineage tracing Auto-documentation Azure-native Human decides

Architecture Overview

📊

Your Power BI Estate

Datasets, reports, dataflows, and the Power BI service, connected directly through the XMLA endpoint and REST API

🧠

Knowledge Layer

Azure AI Search · Azure AI Foundry · model metadata graph with hybrid vector + keyword retrieval

⚙️

Reasoning Engine

Proprietary orchestration framework for DAX parsing, lineage tracing, and natural-language explanation generation

📄

Delivery

Azure Functions · Azure AD · Word / Excel documentation export · Chat interface in your tenant

See It In Action

From an undocumented workspace to full documentation in an afternoon

Watch the Power BI Assistant explore a real semantic model (340 measures, 60 tables), answering analyst questions and generating complete field-level documentation from a customer template.

Core Capabilities

Four capabilities. One semantic model.

Every capability runs on the same underlying knowledge layer. Connect one workspace, then expand to your full BI estate without rebuilding the foundation.

💬

Natural-language exploration

Ask questions the way you'd ask a colleague: "what drove the Q3 margin drop?", "which reports use this table?", "what does this measure actually calculate?" The system understands your model's structure and cites the exact tables and measures behind every answer.

Plain-English queries Cross-report search Business-user friendly Source-cited answers
🔗

DAX & lineage tracing

Every measure and calculated column is parsed, explained, and traced back through its full dependency chain: which tables, which relationships, which upstream measures feed into it. The mystery formula someone wrote three years ago finally makes sense.

Full dependency chains Table & relationship mapping Plain-English DAX explanation Impact analysis before changes
📋

Automated documentation generation

Point it at a workspace and it generates complete documentation for hundreds of fields, measures, and relationships, formatted to your own template. What used to be a two-week manual project becomes a same-day export, kept current as the model evolves.

Customer-template output Hundreds of fields at once Word / Excel export Re-run after model changes
🧭

Report & visual structure understanding

Beyond the data model, the system understands your reports themselves: which visuals exist, what filters are applied, how pages relate to each other, and why a particular chart was built the way it was.

Visual-level explanation Filter & slicer context Page structure mapping Design-intent reasoning
The Interface

Built for analysts,
not engineers.

A conversational interface over your entire Power BI estate that replaces the static wiki and the spreadsheet nobody updates. Built for BI teams, finance analysts, and the business users who depend on reports they didn't build.

  • ✓Ask any question in plain English and get an answer with sources
  • ✓Click any measure to see its DAX, lineage, and explanation
  • ✓Trace a number on a visual back to its originating tables
  • ✓Generate full field-level documentation in one click
  • ✓Export to Word or Excel using your own documentation template
Power BI Assistant: Sales_Analytics_Model (60 tables · 340 measures)
Query: "Why did margin drop in Q3?" 3 measures traced · 2 tables referenced
MEASURE Gross Margin % = DIVIDE([Gross Profit], [Total Revenue], 0)
Depends on Sales_Fact (revenue) and Cost_Fact (COGS), joined on Product_Key and Date_Key. Q3 shows Gross Profit fell 8% while Total Revenue held flat, with the drop concentrated in the DACH region.
Sales_Analytics_Model · Measures.dax · Confidence: High
ROOT CAUSE DACH region: Cost_Fact shows a 14% COGS increase in Q3
Cross-referencing Cost_Fact against Currency_Rates: a portion of the increase is currency-driven (EUR/CHF movement), and the remainder traces to a supplier cost change logged in Q3, not a data-quality issue.
Cost_Fact table · Currency_Rates table · Cross-table reasoning
RELATED Used in 4 reports: Exec Summary, Regional P&L, Board Pack, Sales Ops
This measure is a core KPI referenced across your executive-level reporting. Any change to its definition would affect all four downstream reports, flagged here for awareness before edits.
Cross-report dependency scan · Full impact list available
Getting Started

Running on a real workspace within one week

There's no migration and no changes to your reports. The system connects to your existing Power BI service and Azure environment as read-only.

1

Discovery workshop

We map your Power BI estate, documentation gaps, and the templates you already use. You leave with a concrete pilot scope and proposal.

½ DAY · FIXED FEE
2

Pilot on a live workspace

We connect the system to a real workspace: your actual model, your actual templates. Typically 2 to 3 weeks.

2–3 WEEKS
3

First full documentation run

Your team reviews the first generated documentation set against a live model. We tune the templates and query behaviour together.

DAY 1 OF PRODUCTION
4

Expand & evolve

Connect additional workspaces, add report-level Q&A for business users, or integrate with your BI governance platform.

OPTIONAL RETAINER
Where Manual Documentation Fails

The lineage a data dictionary can't show you, traced automatically.

A static data dictionary tells you a measure exists. It can't tell you that it depends on three tables, breaks silently if a relationship changes, or feeds into the board's headline KPI. The system reads the actual model: the DAX, the relationships, and the dependencies, so the documentation reflects what the model does rather than what someone remembered to write down.

Measure ↔ DAX Table ↔ Relationship Report ↔ Model + cross-report impact

Live example: lineage tracing

Question: → "What feeds into Net Revenue?"
Model says: → DAX references Sales_Fact, Returns_Fact, Discount_Fact
Reasoning: → 3 upstream tables, 2 relationships, 1 nested measure
Result: → TRACED: used in 6 reports, flagged for change-impact

Every trace shows the full dependency chain and where it's used downstream.

Where It's Used

Any workflow where a BI estate needs to be understood

The system adapts to your models and templates, not the other way around.

📚

Documentation backlog

Years of undocumented reports and measures, finally documented fully and consistently in your own template, without weeks of manual work.

🧑‍💼

Onboarding new analysts

New team members ask the model questions directly instead of waiting for the one person who remembers how a report was built.

🔍

Root-cause investigation

When a number looks wrong, trace it back through the full DAX and table lineage in minutes instead of days of manual digging.

⚠️

Change-impact analysis

Before editing a measure or table, see every report and downstream measure that depends on it, so nothing breaks silently.

🏛️

BI governance & audit

Give governance teams a complete, current inventory of measures, sources, and lineage across the entire Power BI estate.

🔄

Migration & consolidation

Understand a legacy or acquired BI estate quickly before migrating, consolidating, or rebuilding it on a new architecture.

Why It Matters

Real impact on BI team productivity and trust

10×
Faster than manual documentation

What used to take a BI team two weeks of manual write-up now completes in an afternoon, across hundreds of fields at once.

100%
Measures traced to source

Every explanation cites the actual DAX and table lineage, with no guessed definitions. Every answer is grounded in the model itself.

∞
Workspaces per deployment

Connect additional workspaces as your estate grows. No schema migrations, no retraining, no vendor involvement to add a model.

Common Questions

Straight answers

Does this modify our reports or datasets?+
No. The system connects as read-only via the Power BI REST API and XMLA endpoint. It reads your model metadata, DAX, and relationships to build its knowledge layer. It never writes back to your reports or datasets unless you ask it to generate a documentation export, and even then that export is a separate file rather than a change to your BI estate.
How does it handle a large or complex estate?+
The serverless architecture scales on demand. We have processed models with hundreds of tables and measures in a single workspace. For very large estates spanning many workspaces, we design the ingestion pipeline to process incrementally with progress tracking, so you see documentation appear as it completes rather than waiting for the whole estate.
Can it use our own documentation template?+
Yes, this is a core design principle. You provide your existing template (Word or Excel), and the system generates output that matches your structure, naming conventions, and level of detail, so it reads like something your team wrote rather than a generic dump of every technical field.
Our model data is sensitive. How is it protected?+
Everything runs within your Azure tenant. The system reads model metadata, including table names, relationships, DAX formulas, and measure definitions, via your existing Azure AD permissions. It doesn't need the underlying row-level data to explain how the model works. We provide a Data Processing Agreement for your procurement and legal review.
We already have a BI Governance Platform. Why do we need this?+
A governance platform tracks policy, access, and lineage at a structural level. This system explains what your model actually means and does, in plain English, for the people using it day to day. The two are complementary: governance answers "who can access this," while the Assistant answers "what does this measure mean, and why does it exist." We can integrate the two if you already use our BI Governance Platform.
How long does a typical deployment take?+
A pilot on a real workspace typically takes 2–3 weeks from kickoff. This gives you a working system on your actual model, with your actual documentation template, not a demo built to look good in a pitch. Full production deployment across additional workspaces, if you choose to proceed, typically follows within 4–6 additional weeks.
Ready to see it working?

See it running on your workspace.
Real data, real answers.

The best way to evaluate this system is a live demo connected to a real workspace from your actual BI estate, or a read-only copy of one. Book a 30-minute call, and we'll walk through a live exploration, then tell you honestly whether your use case is a fit.

Book a Demo Call →

No commitment · No slide deck · Engineers talking through your problem · Based in Switzerland, operating across EU & UK