Before you start
Make sure you have:- At least one warehouse connector in Settings → Connectors — Google BigQuery today
- A semantic layer with models whose source tables that connector can read
Creating a new agent
1
Go to Agents
From the left sidebar, click Agents and then Create agent.
2
Set a name and description
Give your agent a clear, descriptive name — for example, “Paid Search Analyst” or “Campaign Budget Tracker”.
3
Configure the agent
On Agent Builder, tell the agent what it does, which domain it covers, and how to format responses. Optionally include organisation context and business events.
4
Set up data access
Select a warehouse connection, the semantic models (and therefore source tables) this agent may use, and optionally which user groups can open it. Data Access →
5
Save and test
Click Save, then Open chat to test your agent. Adjust instructions and context until the results match your expectations.
Agent settings
Data Access
Warehouse connection, allowed semantic models, and user group restrictions.
Guardrails
Input and output checks: PII masking, off-topic detection, custom rules.
Context
Organisation-wide business, data, and event knowledge agents can opt into.
Testing your agent
Once saved, click Open chat to test your agent. Check:- Does it query the right tables?
- Does it use the correct date ranges and filters?
- Is the output format what you expected?