In this challenge, you will create an agent using Create with AI, test it, and examine the reasoning trace.
Don’t have data yet?
Open the Northwind import guide in the Aura Console and follow the instructions to load the Northwind sample dataset — a food distribution graph with customers, orders, products, and suppliers. The example prompt and questions below use Northwind.
Your challenge
-
Go to Data Services → Agents and create an agent using Create with AI connected to an AuraDB instance with data loaded.
-
Write a prompt that defines the agent’s role, what it can answer, and that it should decline off-topic requests.
-
Save the agent, then ask it at least three questions in the preview panel that you would expect to hit different tools.
-
For each response, expand the Thought section and identify: which tool was selected, what parameters were passed, and why the LLM chose that tool over the others.
-
Review the generated tools against the design principles from the previous lesson: is the scope clear, are the descriptions specific, does each tool map to one question type?
If you are using Northwind, use this prompt:
You are a customer service agent for Northwind Traders, a food distribution company.
Your role is to answer questions about customers, orders, products, categories, and suppliers.
Decline off-topic or harmful requests.And try these questions:
-
Which are the top 5 most ordered products?
-
Who has ordered Pavlova repeatedly?
-
Find products similar to Chai.
Summary
In this challenge, you created an agent with AI, tested it with real questions, and examined the reasoning trace.
In the next module, you will build on this by adding and customising tools with Cypher Template and Text2Cypher.