Microsoft Fabric

Query Lakehouse and Warehouse items from any myai canvas, function, or chat — without copying data out of Fabric.

Data & AnalyticsComing soon

Microsoft Fabric is the corporate datalake / lakehouse / warehouse for a growing set of myai customers. When this integration lands, any Fabric Lakehouse or Warehouse item becomes a queryable source for canvases, functions, and chats — no copy step, no Power BI round-trip.

What you'll be able to do

  • Pull live Fabric data into a chat — ask "what was last month's open-order trend across all brands" and get a table built from Warehouse data without leaving the conversation.
  • Build cross-source dashboards in a single canvas — Fabric tables on one side, BigQuery or Postgres on the other, joined and rendered in the same view.
  • Write back to a Warehouse from automations — a function can update a row, insert a record, or run a stored procedure when conditions fire elsewhere in your stack.
  • Run T-SQL queries with the same shape as Snowflake, Databricks, and BigQuery — one mental model across data clouds.

How it'll work

Connect via Microsoft Entra service principal (or per-user OAuth, depending on whether you want pipeline identity or human-as-themselves access). Behind the scenes, myai uses the Fabric SQL analytics endpoint over ODBC — the same TDS surface that Power BI and external SQL tools use.

Both Lakehouse and Warehouse items are reachable through one credential. Lakehouse exposes read-only T-SQL; Warehouse is full read-write. The integration handles both transparently.

Tell us you want Microsoft Fabric

Use cases we're hearing

  • Cross-brand operational reporting for portfolio companies running multiple ERPs into a unified Fabric layer — Novaria is the lead customer pattern here.
  • Replacing MOR-style PDF reviews with live queries against the Fabric model that backs the report.
  • First-class Epicor-on-Fabric workflows — when Epicor Kinetic data lands in a Fabric Lakehouse, myai reads it natively rather than via the Epicor REST API.

Reference

  • Connection model: T-SQL via ODBC Driver 18 + Microsoft Entra
  • Both Lakehouse and Warehouse items supported through one credential
  • Service principal and delegated OAuth flows both planned
  • Phase 2 will add OneLake DFS access for file-level reads when a customer needs it
PreviousMailchimpNextMicrosoft Outlook