Skip to content

User Defined Prompts

12Port’s AI-Powered Analysis and Insights framework now supports instant, user-defined prompts, allowing users to submit one-time AI queries directly from the interface. This feature provides an interactive experience for exploring and analyzing data without the need to create and save a custom prompt beforehand. Users can enter ad-hoc queries in context and receive immediate insights based on the current dataset.

This section describes how to use the User Defined prompt option and outlines best practices to ensure safe, efficient, and effective AI interactions.

Using User Defined Prompts

User Defined prompts can be used by any user with access to the AI query interface. To submit a one-time prompt:

  1. Navigate to the Report you wish to analyze.
  2. Open the Ask AI option and select User Defined as the prompt type.
  3. Enter your analytical question or instruction in the User Prompt field.
  4. Click Ask AI to submit the prompt along with the current, filtered report dataset.
  5. Once submitted, you will be redirected to the Cloud AI Query Insight report for this submission, to review the Response.

Unlike saved custom prompts, User Defined prompts are one-time executions and are not stored for future reuse. Each query is processed independently and dynamically against the current dataset.

Example Use Case

For example, when reviewing a Jobs report with multiple failed jobs, a user might submit the prompt:

"Why are these jobs failing?"

The AI may return a structured and actionable response such as:

"The jobs are failing consistently because the service on host ubuntu.contoso:22 is unreachable."

This approach enables quick investigation and troubleshooting directly from the interface, without requiring pre-built prompts.

Considerations and Best Practices

When using User Defined prompts, keep the following in mind:

  • Cloud Processing – Queries are processed in your AI Cloud provider; raw report data is transmitted to the AI service for analysis. Ensure that sensitive or restricted data handling policies are followed.
  • Query Scope – Clearly define the question or instruction. Vague prompts may result in incomplete or imprecise answers.
  • Performance Implications – Complex queries or large datasets can increase response time and associated processing costs. Use concise, focused prompts and filtered report datasets where possible.
  • Non-Persistent – Since these prompts are not saved, responses cannot be re-run unless the prompt is re-entered. For repeatable analyses, consider creating a custom prompt.