How the AI Assistant works

Understanding how the Corticon AI Assistant operates will help you get the most from your interactions.

• Prompt Classification: When you submit a prompt, the Assistant first sends your request to the AI service to determine what type of action you want to perform—such as generate rules or explore an asset.
• Context Awareness: Based on the action identified, the Assistant prepares a request that may include your selected rule asset (such as a vocabulary, rulesheet, or ruleflow) and any other information needed to service your prompt.
• Request Processing:
  1. Your prompt is classified by the AI (such as “generate rules” or “explore an asset”).
  2. The Assistant then sends your rule asset and supporting information, along with clear instructions, to your chosen AI provider (OpenAI or Azure OpenAI).
  3. The AI analyzes the provided information and generates a response.
  4. Depending on the type of action, the response will either be displayed in the Assistant panel or an asset will be updated—for example, new rules may be added directly to a rulesheet.
• Stateless Interactions: Each prompt is handled independently—the Assistant does not remember previous interactions, which protects your privacy and ensures clear results.
• Prompt Size Considerations: For very large ruleflows or rule assets, there may be technical limits on the amount of information sent to the AI provider. If your project is especially large, you may get better results by working with individual rulesheets or smaller ruleflows.

Disclaimers

• Policies regarding the handling of data provided to the AI Assistant are between you, the user, and your AI provider. For details, refer to their policies. Progress Software is not responsible for how the AI provider processes data provided through the AI Assistant.

• AI-generated responses may be incomplete or contain errors. Progress Software is not responsible for the accuracy or content of AI-generated responses.

AI Assistant Capabilities by Asset Type

The Corticon AI Assistant provides different capabilities depending on the selected asset type. For some assets, you can ask analytical or explanatory questions to gain insights. For others, you can generate new content such as rules or test cases.

Ask Questions — Enables the AI Assistant to analyze and explain the logic, structure, or behavior of the asset.

Create Content — Enables the AI Assistant to generate new elements for the asset, such as rules in a rulesheet or test cases in a ruletest.

Data Privacy and Usage Notes

Protecting your information is a priority when using the Corticon AI Assistant. Please review the following points to ensure your use of the Assistant aligns with your organization’s standards and expectations:

• Your data and AI provider:

  All information you send through the Corticon AI Assistant is transmitted to your selected AI provider (OpenAI or Azure OpenAI). Data handling, retention, and privacy are governed by your organization’s agreement with that provider. Be sure to follow any internal guidelines for sharing business or sensitive data.

• No retained memory:

  The Assistant does not retain any information between prompts. Each request is independent and processed without knowledge of your previous interactions.

• Prompt size and project limits:

  For very large ruleflows or rule assets, there may be token limits on the amount of data that can be sent to the AI provider in a single request. If you encounter issues, try working with individual rulesheets or smaller ruleflows.

• Validation is required:

  Because results can vary, you are responsible for reviewing and testing all AI-generated rules and suggestions before using them in production or critical business processes.

• Results may vary:

  AI-generated content is meant to assist and accelerate your rule modeling work, but it may not always be perfect or fully aligned with your business logic. Always use your professional judgment when evaluating responses.