Using the DataStage Assistant
The DataStage® Assistant is an AI-powered tool that is designed to streamline your interaction with the IBM DataStage UI. It enables users to build data integration pipelines, interact with the canvas, and access documentation by using natural language.
The Assistant is available today on DataStage on IBM Cloud Pak® for Data as a Service within the Dallas, Frankfurt, and Sydney data centers.
- Design DataStage flows,
- Get answers from DataStage documentation,
- Generate Transformer expressions,
- Understand and interpret Transformer expressions.
The Assistant uses built-in actions and is trained on official IBM DataStage documentation to deliver highly accurate and reliable responses.
Starting the DataStage Assistant
- From your project home page, select the Manage tab.
- Within the left sidebar, select DataStage.
- Go to the AI tab from the main menu.
- Click the toggle so that it shows On - you have now enabled the DataStage Assistant.
- Go back to the Assets main tab.
- Open up a DataStage flow canvas (either by opening an existing DataStage flow or selecting New Asset > Transform and Integrate Data).
- Click the DataStage Assistant icon
.
- Select an action from the main menu and then follow the instructions.
What can the DataStage Assistant do?
The DataStage Assistant enables completion of the tasks that are outlined in the following table.
| Action | Description |
|---|---|
| Build a flow | You can tell the Assistant about the main data sources and targets that you want to connect to. You can also tell the Assistant about the main actions you want to take. The Assistant predicts the different connectors and transformations stages that are required and outputs them to the canvas. |
| Modify/Add/Remove stages | You can rename a stage, add a stage, remove a stage, or configure existing stage properties within a DataStage flow. When you want to rename stages, click the AI button for the Assistant to automatically generate stage names based on stage properties. |
| Rename your flow | You can change the name of your existing DataStage flow. |
| Transformer expression | You can learn how to use the Transformer Stage with the DataStage Assistant. For more information, see: Transformer stage in DataStage. You can:
|
| Ask a DataStage question | You can ask the Assistant anything about DataStage and it generates an answer based on its knowledge of the product. The Assistant is trained on the official IBM DataStage documentation. |
Troubleshooting the DataStage Assistant
- Rephrase the prompt. The assistant might not understand your phrasing. Try to use different wording, if possible.
- Restart the conversation. The assistant might be getting confused by other things you discussed with it. Click the refresh button in the upper right of the chat interface to start over.
- Be sure that you are asking about DataStage. If you are asking about something else, it might be outside of the assistant's capabilities.
- Try asking about something specific. If your question is too broad, the assistant is more likely to return results that are not relevant to your request.
- If the assistant lacks functions for a feature or task you find valuable, you can open an enhancement request in IBM Data and AI Ideas Portal for Customers.
Data Privacy with the DataStage Assistant and opting out
When you interact with DataStage Assistant, IBM watsonx Assistant processes your questions and requests by using metadata and data from your DataStage instance. Only the text that you input into the assistant gets processed by watsonx Assistant. None of your actual data or metadata leave your DataStage instance at any time.
- From your project home page, go to the Manage tab.
- In the left navigation bar, select DataStage.
- In the main navigation bar, click the AI tab.
- Locate the DataStage Assistant toggle and switch it to the Off position.