PADU Income Intelligence Agent

Unit PADU, Kementerian Ekonomi Malaysia

Agentic AI powered by Google Gemini — data from data.gov.my

Use Case

Unit PADU, Kementerian Ekonomi Malaysia uses this agent to analyse household income vulnerability using live open data from data.gov.my. The agent can answer questions in natural language such as:

• What is the B40 household income threshold?
• Fetch the latest household income data
• Fetch labour force unemployment statistics
• Train and evaluate the ML classification model
• Analyse which income group needs policy intervention

The agent autonomously picks the right tool, calls the data.gov.my API, and returns a policy-ready answer — all without any coding.

Step 1: Google AI Setup

Configure Google Gemini authentication.

1. Credentials Configuration — API key is pre-filled. Just run.
2. Google AI Studio Authenticator — validates the key with Google.
3. Gemini LLM Selector — uses gemini-2.5-flash model.

Step 2: Tool Registration

These nodes load the 4 PADU tools the agent can call:

• ToolFetchIncome — pulls household income data from data.gov.my API
• ToolFetchLabour — pulls unemployment/labour force data
• ToolTrainModel — trains B40/M40/T20 classifier
• ToolAnalysePoverty — returns policy summary and risk flags

List Files/Folders scans the Tools folder. Workflow to Tool converts each into a callable AI tool.

Step 3: Agent Chat

The Agent Chat View is the live chat interface.

To use:

1. Press the green Run button (▶) to execute all nodes
2. Right-click Agent Chat View node
3. Select Open View
4. Type your question and press Enter

Try: "Fetch the latest household income data and tell me the B40 threshold"