Configuring generative AI
Generative AI (GenAI) capabilities improve productivity and simplify complex tasks. This feature integrates with IBM® watsonx® to provide contextual guidance and automation.
Before you begin
12.2.x and later This feature is available in 12.2.x and later versions of Guardium®.
Obtain watsonx subscription credentials (URL, Project ID, API Key).
Ensure that you have administrative access to the Guardium central manager (CM).
Depending on your Guardium version, download and apply the appropriate patch from Fix central:
12.2.1 Gen AI patch 12.0p80000.
12.2.2 Gen AI patch 12.0p80003.
About this task
This procedure sets up a GenAI app node on the CM. After you enable GenAI in your Guardium environment, you can use it with any Guardium feature that supports GenAI.
Procedure
- Deploy a fresh standalone collector.
- Register the newly deployed collector with the CM.
- Run the following CLI command: store unit type app-node
- From the CM, go to .
- Apply the Gen AI patch for your Guardium version.
- Under Connection settings, enter your watsonx credentials.
- After you test and save the connection. switch the Enable watsonx toggle to Yes. On the Central management page, the newly added unit will appear as an App-Node unit type.
Results
After you enable GenAI in your Guardium environment, you can use it with any Guardium feature that supports GenAI.
What to do next
12.2.x and later Use GenAI with SOX ticket reconciliation.
SOX ticket reconciliation automates mapping of database changes to ServiceNow® tickets, reducing audit time and improving accuracy.
You can import change tickets by creating a custom table from the relevant datasource. Enable the Calculate embeddings option to create a new column type in your custom table. Map the columns from the custom table to the corresponding entries from the drop-down list in the new column. This allows the system to perform AI-driven matching.
Next, set up an audit process and enable the Change ticket matching option, linking the audit to the custom table containing ticket data. When the audit runs, the AI ticket reconciliation view displays results, offering two approaches: manual mapping, where you can review the top five suggested matches for each change, or automatic mapping, which assigns the best matching ticket to all changes in one step. Automatic mapping can reconcile in seconds, saving hours of manual effort. You can also override or adjust mappings if needed, ensuring flexibility and accuracy throughout the process.