Skip to content

BRK298 Breakout Wed 18 Nov 13:45 PST

Bringing conversational analytics to where decisions happen

Amir Jafari, Sara Vredevoogd

Summary

From the recording

Microsoft's Fabric IQ team said tools built on Power BI semantic models are generally available in Copilot Chat, Cowork, MCP and GitHub Copilot, and cited more than 22 million semantic models used daily. Data agent in Copilot Chat is to be generally available soon. Fabric apps general availability was announced this week; building apps from data agents and A365 integration start rolling out in the next few weeks. OptiGuide is in private preview, with public preview soon. Up to 1000 rows is now available in the standard runtime. No dates, pricing or licensing were given.

Session description

Microsoft's catalogue description

Data should meet users where decisions happen. Learn how Microsoft Fabric brings conversational analytics into the tools people use every day. Powered by Fabric IQ, which unifies OneLake, the Power BI semantic layer, ontologies, and data agents, Fabric provides a trusted intelligence layer for copilots, applications, and workflows. See how organizations connect analytics, context, AI, and governance to turn questions into actions.

Demo timestamps

Microsoft's player does not support links to a specific time, so open the recording and skip to the time shown.

  1. Skip to 00:05:40

    Switching to Cowork for longer-running follow-up work

    After reviewing Power BI data on whether a Miami move makes sense, the speaker switches to Cowork. It gets the same Power BI data and Work IQ context and is asked to do longer-running follow-up work.

  2. Skip to 00:11:44

    Excel with Fabric IQ tools and a pivot table

    In Excel, Fabric IQ tools are @-mentioned to ask about IT spend variance to plan in Europe. Copilot is then asked to build a pivot table.

  3. Skip to 00:12:26

    Consent for external data connection to semantic model

    Copilot prompts for consent to create an external data connection to the semantic model, so later refreshes follow the data team's model.

  4. Skip to 00:23:03

    End-to-end recap across multiple data sources including a lakehouse

    Recap of the end-to-end pass across multiple data sources, including a lakehouse, and then building on the result.

  5. Skip to 00:25:20

    Switching to GitHub Copilot with a copied link

    The demo switches to GitHub Copilot using a copied link. It returns to GitHub Copilot at 00:26:44.

Watch on-demand (opens in a new tab)

← All sessions