BI Pixie Workload for Microsoft Fabric is Now Generally Available

Last month we announced the public preview of BI Pixie running natively inside Microsoft Fabric. Today we are taking the final step.

BI Pixie Workload for Microsoft Fabric is now generally available in the Fabric Workload Hub, having met Microsoft’s workload publishing requirements for GA. Any Fabric tenant can add BI Pixie directly from the BI Pixie page in the Workload Hub and start tracking how Power BI reports are actually used without leaving Fabric.

A first-class Fabric workload

Run this way, BI Pixie is a native Fabric item that lives in your workspace, right next to your lakehouses, semantic models, and reports. There is nothing separate to log into: no portal to manage, no infrastructure to deploy, and no separate user directory to maintain. Once a tenant admin adds BI Pixie from the Workload Hub, anyone in the tenant can create a BI Pixie item straight from the workspace + New menu.

Because it runs inside Fabric, BI Pixie inherits the platform you already trust. Access is governed by your Fabric workspace roles, so the people who can edit a report are the people who can track it. And the BI Pixie Dashboard is set up automatically in the same workspace, so you are analyzing adoption and engagement within minutes of adding your first Pixies.

What’s new at GA

The public preview brought the full engagement experience (managed reports, the BI Pixie Dashboard, embedded surveys, and RLS auditing) into Fabric. GA adds the capabilities that make BI Pixie ready to adopt across your organization.

Scalable telemetry processing

Popular reports can generate a lot of engagement events, and that volume has to be turned into the numbers behind your BI Pixie Dashboard quickly and affordably. Earlier releases loaded the data with an import-mode refresh, which reprocessed everything on each run and did not scale as telemetry grew. GA moves this to a Fabric notebook that runs on your own capacity, processing your engagement data incrementally so it keeps up with high volumes at a fraction of the cost. The result is the same dashboard you already know, refreshed faster and ready to grow with your organization. The same notebook approach powers the customer-run maintenance notebooks that you run under your own identity for data retention and user erasure. Learn more about BI Pixie notebooks.

Data residency in your OneLake

By default, your raw engagement events are stored in a dedicated, secured container in the BI Pixie cloud, in your capacity’s region. Your BI Pixie item reaches that data through a shortcut in your BI Pixie Lakehouse, and the daily BI Pixie ETL run loads new events into your dashboard tables. Your processed dashboard data always lives in your own Fabric workspace.

With Data residency, the raw events move to your tenant too. The BI Pixie cloud still collects the engagement events from your users, but instead of keeping them, it pushes each new event to a Lakehouse in your own Microsoft OneLake. Nothing is retained in the BI Pixie cloud. This keeps your telemetry inside your Fabric tenant, governed by your own capacity and tenant policies, which is what teams with data-sovereignty or regulatory requirements need. You can switch it on or off at any time. Read more.

Bring Your Own Storage (BYOS)

BYOS, an Enterprise capability, takes this one step further: you host the BI Pixie telemetry collector yourself, in your own Azure subscription, so events go straight into an Azure Data Lake Storage Gen2 account you own and never touch BI Pixie infrastructure at all. You keep the full Workload experience to manage reports, add Pixies, and install the dashboard, while your raw telemetry stays entirely within storage you control.

AI assistants

Engagement data is only as valuable as the questions you can ask of it. AI Assistants let people and agents explore report engagement, adoption, and user satisfaction in plain language, grounded in the same data as your BI Pixie Dashboard, through two options:

• Fabric Data Agent: a native, zero-install chat agent for everyone in your workspace. Ask in plain language inside Microsoft Fabric and Microsoft 365 Copilot.
• BI Pixie MCP Server: connect your own AI tools (Claude, GitHub Copilot, Codex, Azure AI Foundry) for exact, repeatable, programmatic access.

But the bigger opportunity is what BI Pixie brings to every agent you already run. When a data agent turns a question into a query against a large model, it has to guess which table, column, and value the user means, and that guesswork is where agents go wrong. BI Pixie removes the guesswork by grounding your agents in how people actually use your reports. It knows which fields users filter on most, which values they select, and which reports and pages carry the business, so your agents resolve ambiguity the way your users would and focus on the data that matters most to the people reading the dashboards. Point a Fabric Data Agent or your own MCP client at your BI Pixie model, and it answers not just from your data, but from your organization’s real behavior.

The BI Pixie MCP Server is generally available too. Instead of writing DAX or exporting tables, you ask which reports are unused and could be retired? what is my stickiness? which pages get the most clicks? The answers come straight from your BI Pixie semantic model, match your BI Pixie Dashboard exactly, and the server is strictly read-only.

Start free

Start on the Free tier: up to 3 tracked reports and 2,000 billed sessions, with full access to Pro and Enterprise features so you can evaluate everything, and no credit card required. When you are ready to grow, a single BI Pixie subscription (through the Azure Marketplace or the BI Pixie subscription portal) covers every BI Pixie item across your tenant. And on Enterprise, you can divide that one subscription into isolated licenses for separate business units or cost centers, each with its own report quota, data isolation, and chargeback.

Get started in five minutes

1. Tenant admins: add BI Pixie from the Workload Hub. The Add BI Pixie to Your Tenant guide walks through the tenant setting, distribution scope, and verification.
2. Report owners: create a BI Pixie item, add Pixies to a report, and watch the events flow. The Getting Started guide covers your first five minutes.

Open BI Pixie in the Fabric Workload Hub

Prefer Cloud or self-hosted? BI Pixie meets you wherever your Power BI runs. See the deployment options for every way to run it.