VantegrateNative API integrationIntegrate Power BI with AI agentsand ask your dashboards questions in natural language
Metrix, Vantegrate's data agent, connects to your Power BI through the official API (Power BI REST with Execute Queries): it turns your question into DAX, runs it against your semantic model and gives you the answer, without you opening the dashboard. The data stays in your Power BI; the agent queries it but never moves it.
How the data flows
What does it mean to integrate Power BI with Vantegrate's AI agents?
Integrating Power BI with Vantegrate's AI agents means that Metrix queries your data through the official Power BI REST API: when you ask something like how this month's revenue is tracking, it turns the question into DAX, runs it with Execute Queries against your semantic model and gives you the answer in seconds, without opening the dashboard. It reads your own model, not a separate copy.
That is different from a bot that lives outside your stack with an outdated spreadsheet. Here the agent queries the same semantic model that feeds your dashboards (a single source of truth), so its answers match what your team sees. The connection uses OAuth 2.0 with a Microsoft Entra ID service principal, with read permissions on the models you authorize; your data stays in your Power BI.
What do you want to connect to your Power BI?
Pick what you need and we'll write the message for you.
1 · Pick what to solve
2 · What our team receives
Vantegrate
online
What data Metrix reads from your Power BI
The agent works on standard Power BI objects in read mode: it queries your semantic models with DAX to answer, without touching or altering the source data.
Semantic models and measures
The agent queries your semantic models (Power BI datasets) and reuses the DAX measures you've already defined, so the answer uses the same calculation logic as your dashboards.
Questions in natural language
Your team asks in plain English and Metrix turns the question into a DAX query, runs it with Execute Queries and returns the figure, with no queries to write and no need to open the Power BI service.
Only what you authorize
You define which workspaces and models each integration queries. The service principal accesses them with read permissions, no more and no less than the use case requires.
How the integration works, step by step
Vantegrate handles the setup and rollout. Your tenant admin authorizes it, and your team validates it and starts using it.
Service principal in Entra ID
We register an app in Microsoft Entra ID (Azure AD) with a service principal (app ID and secret) and a security group, for a server-to-server connection with no user credentials.
Tenant and workspace permissions
The admin enables service principals in the admin portal and turns on the Execute Queries setting; we add the service principal to the workspace with read permissions on the models.
DAX queries
Metrix turns each question into DAX and runs it with Execute Queries against the semantic model, reusing your measures so the number matches the dashboard.
Validation with real data
We test with your models and your real KPIs, validate the answers with your data and IT teams, and roll out in phases.
Execute Queries vs. exporting reports by hand
Power BI exposes an official REST API and, within it, the Execute Queries operation, which runs a DAX query against a semantic model and returns the data in read mode. That's the path Metrix uses to answer: instead of exporting a report to Excel and recalculating, the agent asks the same model that feeds your dashboards and gets the current number.
The difference from exporting by hand, or from a bot that copies data, shows up in consistency. The agent reuses your existing DAX measures, so the answer uses exactly the same logic as the dashboard and no third version of the KPI shows up. Execute Queries supports only DAX (not MDX or DMV queries), which is why the integration is designed around the measures and tables in your model.
| Dimension | Exporting by hand / external bot | Vantegrate's API integration |
|---|---|---|
| Data source | An exported copy that drifts | Your semantic model, live |
| Calculation logic | Recalculated separately, may differ | Your DAX measures, the same as the dashboard |
| Connection | Manual or user credentials | OAuth 2.0 with a service principal |
| Access | Broad or uncontrolled | Read permissions on what you authorize |
| Language | Spreadsheet formulas | DAX via Execute Queries |
Comparison of patterns; the exact scope is defined with your team based on your tenant, your workspaces and your models.
Your data stays in your Power BI: Metrix queries your semantic model through the API, in read mode; it doesn't take your database anywhere else or alter the source data. The connection uses a service principal with read permissions, and every query is logged and auditable.
What your tenant admin needs to enable
Connecting an agent to Power BI isn't self-service: it requires action from the tenant admin. They need to register an app in Microsoft Entra ID, turn on the admin portal setting that allows service principals to use Power BI APIs, enable the Execute Queries tenant setting and add the service principal to the workspace with read and build permissions on the models.
One detail that prevents hard-to-diagnose errors: for a service principal, Microsoft recommends not adding delegated API permissions in the Azure portal, because they aren't used and tend to break the connection. The integration also respects the API limits (for example, the per-minute query quota), with scoped queries and retries. And according to Microsoft, Execute Queries doesn't support service principals on models with row-level security (RLS) or with SSO enabled; in those cases, the connection path is defined with your team before we start.
About the API limits: Power BI throttles usage by frequency and responds with an HTTP 429 when you go over (for example, 120 queries per minute per user); the integration is designed to stay within those quotas.
You've seen how it connects. Want to walk through your case?
Tell us how your Power BI is set up and we'll tell you what data we need and where the agent connects.
An agent connected to your Power BI vs. a standalone tool
| Standalone AI tool | Agent connected to your Power BI | |
|---|---|---|
| Data source | A copy that drifts | Your source of truth, live |
| Where your data lives | In a third-party system | In your environment, encrypted end to end |
| Result of each question | A one-off export | An answer from live data, where your team works |
| Permissions and audit | Broad and hard to audit | Read-only role and a log of every query |
| Reaching production | Often stalls as a pilot | Validated, phased rollout |
Why add a conversational channel on top of your dashboards
Third-party figures, each with its published source, to size up the challenge. None of them is a Vantegrate result.
25-35%
Of employees use BI tools, depending on the study: 25% for BARC and Eckerson Group, 35% for Gartner
95%
Of generative AI pilots never reach production with measurable business impact
The BI adoption figure is an industry benchmark, not Vantegrate data, and it varies by source. The MIT NANDA figure reflects the gap between experimenting with AI agents and running them in production.
Which agent runs on your Power BI
The data integration is used by Metrix, the conversational BI agent: it works on your semantic model in real time and answers with the same logic as your dashboards.
The integration doesn't take your data anywhere
The principle is simple: AI comes to your data, not your data to the AI. Metrix queries your Power BI through the official API, in read mode and with a scoped service principal, and the agents run on certified cloud infrastructure.
- Connection through the official Power BI REST API with OAuth 2.0 and a Microsoft Entra ID service principal, with no scrapers and no fragile bridges.
- The agent queries your semantic models with DAX in read mode; your database and your models stay in your Power BI.
- Access scoped by read permissions on the workspaces you authorize, with traceability for every query.
- The agents run on Salesforce or Oracle Cloud Infrastructure, whose SOC 2 and ISO 27001 certifications belong to those platforms, not to Vantegrate or Power BI.
Frequently asked questions about the Power BI integration
What data and IT teams usually ask before connecting an AI agent to their Power BI.
How do I integrate Power BI with an AI agent?
How do I integrate Power BI with an AI agent?
With a connection through the official Power BI REST API, using OAuth 2.0 with a Microsoft Entra ID service principal. Metrix turns your question into DAX, runs it with Execute Queries against your semantic model and returns the data in read mode. Your tenant admin enables service principals and the Execute Queries setting and adds the service principal to the workspace; your team validates and starts using it. Meet the data agent at Metrix.
What language does it use to query my Power BI data?
What language does it use to query my Power BI data?
DAX. The API's Execute Queries operation supports only DAX (not MDX, DMV or INFO functions), so Metrix turns your question into a DAX query and reuses the measures you've already defined in your semantic model. That's why the number it returns matches your dashboard: it uses the same calculation logic.
Can you query my data without touching the dashboard?
Can you query my data without touching the dashboard?
Yes. Execute Queries works in read mode on the semantic model: the agent queries the data and returns it without opening or modifying the dashboard, and without altering the source data. The Execute Queries tenant setting needs to be enabled, and the service principal needs read and build permissions on the model.
What does the tenant admin need to enable?
What does the tenant admin need to enable?
It isn't self-service. The admin registers an app in Microsoft Entra ID (a service principal with an app ID and secret), turns on the Power BI admin portal setting that allows service principals to use the APIs, enables the Execute Queries setting and adds the service principal to the workspace with read permissions. Microsoft recommends not adding delegated API permissions in Azure, because they aren't used and tend to cause errors.
Do I need the Power BI gateway?
Do I need the Power BI gateway?
Only if the source data is on premises and Power BI needs to refresh it through the gateway. To query a semantic model that's already published to the Power BI service through the API, you don't need a gateway: Metrix asks the model in the cloud directly.
Does it work if you use Power BI inside Microsoft Fabric?
Does it work if you use Power BI inside Microsoft Fabric?
Yes. Microsoft Fabric is Microsoft's analytics platform, and Power BI is one of its workloads, alongside Data Factory, Data Warehouse and Real-Time Intelligence, as the Fabric documentation details; your workspaces and your semantic models stay the same, according to Microsoft. Metrix queries them the same way: with DAX through Execute Queries, with the service principal and read and build permissions on each model.
Metrix, the data agent
How your team asks about its KPIs in natural language and gets the answer from your models, without opening a dashboard.
Learn moreAll integrations
The full map: the BI tools, ERPs, CRMs and other tools the AI agents connect to.
Learn moreAI agents for enterprises
The map of the Vantegrate suite: sales, marketing, documents, data and logistics.
Learn moreLet's connect AI agents to your Power BI
Tell us which semantic models and KPIs you have in Power BI and we'll show you what the agent queries, which service principal we request and what asking about your metrics in natural language looks like. A 30-minute conversation, no commitment.
Francisco Morales, co-founder, takes your call. We reply on WhatsApp within 4 business hours, no strings attached.
