Build the Web IQ lab foundation

In this exercise, you prepare the internal-data side of the Zava Retail solution. You create a Fabric workspace and lakehouse, load synthetic retail data with a notebook, and publish a Fabric data agent that answers inventory questions in natural language.

This exercise should take approximately 45 minutes to complete.

[!NOTE] Fabric data agents require a paid F2 or higher Fabric capacity or Power BI Premium P1 or higher. Your tenant administrator must also enable the required cross-geo AI processing settings.

Prerequisites

To complete this exercise, you need:

  • Access to Microsoft Fabric with permission to create workspace items.
  • A workspace assigned to a supported Fabric capacity.
  • Permission to create and publish a Fabric data agent.
  • The files in this lab's Labfiles folder.

Create a Fabric workspace

  1. Open Microsoft Fabric and sign in.
  2. In the navigation pane, select Workspaces, and then select New workspace.
  3. Enter Web IQ - Zava Retail followed by your initials to make the name unique.
  4. Expand Advanced, and assign the workspace to a supported Fabric capacity.
  5. Select Apply.

Create the Zava lakehouse

  1. Open your new workspace, and then select New item.
  2. Search for and select Lakehouse.
  3. Enter ZavaRetailLakehouse for the name.
  4. Leave Lakehouse schemas selected, and then select Create.
  5. In the lakehouse Explorer, confirm that Tables and Files appear.

Upload the retail data

  1. In the lakehouse Explorer, select the ellipsis next to Files, and then select New subfolder.
  2. Name the folder zava-data.
  3. Select the ellipsis next to zava-data, and then select Upload > Upload files.
  4. Upload products.csv, inventory.csv, and carrier_routes.csv from Labfiles/data.
  5. Confirm that all three files appear in the folder.

Load the Delta tables

  1. Return to the workspace, select Import > Notebook > From this computer, and import Labfiles/notebooks/load-zava-retail-data.ipynb.
  2. Open the notebook, and select Add lakehouse in the Explorer pane.
  3. Select Existing lakehouse, choose ZavaRetailLakehouse, and select Add.
  4. Confirm that the lakehouse is marked as the default lakehouse.
  5. Select Run all, and wait for the Spark session and all cells to finish.
  6. In the lakehouse Explorer, refresh Tables and confirm that products, inventory, carrier_routes, and stock_status appear.

The final notebook output lists products in projected-stock order. Products with negative projected units have a High stockout risk.

Create the Fabric data agent

  1. Return to the workspace, select New item, and search for Data agent.

  2. Select Data agent, enter Zava Inventory Data Agent, and select Create.

  3. Select Add a data source, choose Lakehouse, and select ZavaRetailLakehouse.

  4. Add the products, inventory, carrier_routes, and stock_status tables.

  5. In Data agent instructions, enter:

    Answer questions about Zava Retail products, inventory, stockout risk, and carrier routes.
Use stock_status for stockout questions. projected_units equals on_hand_units plus
inbound_units minus expected_weekly_sales. A negative projected_units value means
high stockout risk. Do not infer causes or use information outside the selected tables.

  6. Add this example question and query if the experience offers Example queries:

    Question: Which products are at risk of stockout?
Query: SELECT product_name, distribution_hub, on_hand_units, inbound_units,
expected_weekly_sales, projected_units, stockout_risk
FROM stock_status
WHERE stockout_risk IN ('High', 'Medium')
ORDER BY projected_units ASC;

Test and publish the data agent

  1. In the data-agent chat, enter Which products are at risk of stockout?.
  2. Confirm that the response uses stock_status and identifies products with high or medium risk.
  3. Enter Which carriers serve the Memphis distribution hub?.
  4. Confirm that the response uses carrier_routes and lists the carriers.
  5. Select Publish, provide a description such as Initial Zava inventory agent, and confirm the publication.
  6. Record the published data-agent name and workspace name for Exercise 2.

[!IMPORTANT] Foundry uses the signed-in user's identity to call the Fabric data agent. Each learner must retain access to the published data agent and its underlying lakehouse.

Summary

In this exercise, you created the Fabric workspace and lakehouse, loaded the Zava Retail data into Delta tables, and published a Fabric data agent for governed internal-data questions.

You have successfully completed this exercise.

Clean up

Keep these resources for the remaining exercises. After you complete the full lab, delete the Fabric workspace if you no longer need it.