Discovery AI

New! Enterprise Discovery AI

Discovery AI brings clarity to complexity. It simplifies DAX, Power Query and SQL across your semantic models, creating transparency and trust in your Power BI landscape and enabling confident governance at scale.

Illustration of five people sitting around a table in discussion, surrounded by multiple charts and graphs with warning icons and exclamation marks. The reports show numerical data, bar charts, and pie charts, but include error symbols and caution signs, indicating issues and complexity in interpreting the data. A question mark above one person suggests confusion or uncertainty about resolving the problems.
AI That Builds Transparency, So You Can Build Trust

What is Discovery AI?

Discovery AI is the latest innovation from Power BI Sentinel, designed to help business leaders gain clarity and trust in their Microsoft Power BI estate.

Powered by Large Language Model (LLM) technology, it translates complex technical logic into understandable, audit-ready insight that supports confident governance at scale.

Why use Discovery AI?

Power BI estates grow quickly, and the logic behind them becomes increasingly complex. DAX formulas, Power Query steps and SQL statements are often buried within semantic models, making it hard for teams and leaders to understand how reports work or why metrics behave the way they do.

 

Discovery AI removes this complexity by:

Simplifying Complex Code

Converts DAX, Power Query, and SQL into human-readable explanations, broken down into clear, bullet-pointed steps.

Providing Actionable Insights

Suggests improvements to existing code for better performance and maintainability.

Highlighting Business Context

Generates potential business cases for measures and calculations, helping stakeholders understand their purpose and impact.

How It Works

Metadata Processing

Discovery AI scans your Power BI estate, analyzing semantic model metadata, source queries and measures

BUSY_DATA_GRAPHIC@2x
Data_Cleaning_Robot_DiscoveryAI@2x

AI Powered Interpretation

Using LLM technology, it translates technical syntax into plain language explanations.

Transparency Delivered

End users gain clarity on:

  • How items are built step by step
  • High level summaries
  • Possible Use Cases
  • Potential performance improvements
Illustration of five people sitting around a table in discussion, surrounded by multiple charts and graphs with success and security icons. Table showing agreement and successful outcomes

The Result: Trust and Confidence

Even with complete access to code, most teams struggle to understand how DAX, SQL and Power Query logic shapes their reports. When the meaning is unclear, trust in metrics suffers and decision-making slows.

Discovery AI transforms that code into plain-language explanations. It builds shared understanding, strengthens governance and gives every user confidence in the reports they rely on.

Screenshot of a Power BI Sentinel's Discovery AI interface showing a DAX code snippet under the section “Measures referencing this table.” The header indicates “Owner Table: All Measures” and “Found 8 References.” The code block is titled “Suggested Action” and includes variables such as WorkspaceID, User, and AccessRequired, followed by conditional logic using IF, SWITCH, and FILTER statements. The code suggests actions like removing users from workspaces or downgrading roles based on permissions and activity. A “Discovery mode” badge appears at the top right of the code block.
Power BI Sentinel's AI documentation summary page explaining a DAX expression for managing Power BI workspace access. The page includes sections titled “Summary,” “Description,” “Functionality,” “How it works,” “Optimization,” and “Possible use cases.” Key points describe logic for removing, downgrading, or leaving user or AD group access unchanged based on activity and role rank. Functionality lists variables like WorkspaceID, User, and AccessRequired, and explains checks for user activity, group permissions, and role downgrades. Optimization tips include storing filtered tables and using conditional checks. The page uses highlighted keywords and bullet points for clarity.

New Plugin - Coming soon!

Confluence - As part of Discovery AI

Coming soon to the Discovery AI license tier. Compatability with Atlassians Confluence platform. Utilising Confluence you will be able to combine Discovery AI's data outputs with your existing estate documentation, and allow your business users to consume it via the confluence platform.

Following this, we plan to release a future plugin for Microsoft Teams!

Power BI Sentinel's Confluence interface showing the Sentinel Governance Report 2.10 dataset documentation page. The top section includes navigation breadcrumbs and options such as “Semantic model explorer,” “Advanced documentation,” and “Generate PDF documentation.” Below, two main panels display: “Reports Using This Dataset” listing the report name and workspace, and “Data Sources & Connections” showing an SQL server and database details. Additional panels show “Refresh Schedule” with settings like time zone (UTC) and mail-on-failure notifications, and “Refresh History” with a table of recent refresh attempts, including status indicators like “Completed.”