Step 4: Build Query with Azi

Goal: Create a Warm Query and build it with Azi in one governed conversation.

---

What is a Warm Query?

A Warm Query is the heart of the Control Plane:

• Named, saved queries against your data
• Each becomes an API endpoint automatically
• Queries are governed - Azi proposes, you approve
• Full audit trail of query creation and changes

The query doesn't run constantly. It runs when called - "warm" and ready.

---

Option A: Ask Azi to Create and Build

The recommended approach - one conversation creates and populates your query:

1. Make sure you're inside your Surface (BrewMonitor)

2. Open the Azi chat panel on the left

3. Type a complete request:

   "Create a warm query called temperature-avg that shows average temperature over the last hour"

4. Azi proposes both the query container AND the KQL:

5. Review the proposal, then click Approve

---

Option B: Create Manually, Then Build with Azi

If you prefer step-by-step control:

Step 1: Create the Query Container

1. Inside the Surface, find Warm Query in the Node Bank
2. Drag it onto the canvas
3. Name it (e.g., "temperature-avg")
4. Connect data to it from your connection

Step 2: Connect Data to the Query

1. Draw an edge from your connection node to the warm query node
2. The edge automatically labels as "queries" - this shows the data relationship

Step 3: Open the Query Manager

1. Click on the Warm Query node
2. In the Inspector, click to enter the Query Manager
3. This opens a modal with the query editor and Azi panel

Step 4: Ask Azi to Build the Query

In the Query Manager Azi panel, type:

"Show me the average temperature over the last hour"

Azi proposes the KQL, you review and approve.

---

Review the Proposal

Before approving, you can:

• Read the KQL - Understand exactly what will run
• See the reasoning - Azi explains why she chose this approach
• Test it - Click Run Query (coral button, bottom of editor) to preview results
• Edit - Modify the KQL directly if needed

---

Approve, Reject, or Edit

Three choices:

Action	Button	When to Use
Approve	✓ checkmark	The query does what you want
Reject	🗑️ trash	Start over with a different request
Edit	✏️ pencil	Minor tweaks to the proposed KQL

Click the checkmark (✓) to approve and save the query.

---

The Audit Trail

Every interaction is logged:

What	Recorded
Your request	"Create a warm query called temperature-avg..."
Azi's proposal	The KQL and reasoning
Your action	Approved by [your identity]
Timestamp	When it happened

---

Commit and Deploy

Save and deploy your query the same way you did in Step 2:

1. Click the Save icon (upper-left)
2. Click Deploy

Your query is now live and API-accessible.

---

Verify

Your query is complete when:

• KQL appears in the query editor
• Status shows "Ready" or "Query saved"
• Run Query returns expected results
• Changes are committed and deployed

---

What You Built

---

What's Next

Your governed query is live. In the next step, we'll access it via API - the endpoint is ready and waiting.

Next: Step 5 - Use Your API →