Configuring the GenAI Connect Rule for Pega Knowledge

Pega added a new feature to Pega Knowledge in version 23. This feature is called Assisted authoring with Pega GenAI™.

These user-definable AI-generated options in the Pega Knowledge Authoring Portal aid authors in creating new content or existing content, including grammar checking and index creation, when a prompt manager creates prompts for these tasks.

A single click provides content authors with AI-generated assistance for the article titles, abstracts, and content bodies. The new AI-assisted authoring option can streamline content creation while maintaining organizational consistency.

Here is a screenshot with some sample options created by a prompt manager.

If you are on a Pega Cloud instance and have a Pega GenAI Connect license, you can enable and configure this option.

Here are the steps:

Log in as a user with the KMSysadmin role. Go to App Studio and then choose Configurations. Go to the Configuration Set for Knowledge Management - AI settings, and set Assisted authoring with GenAI to true.

To view or change the GenAI model used in Pega Knowledge, in App Studio, select AI Designer > GenAI Connect. Open the rule to view the model as shown:

More information on Pega Knowledge Assisted authoring with Pega GenAI™ is located here: Pegasystems Documentation

More information on Configuring advanced GenAI Connect settings is located here: Pegasystems Documentation

More information on the Pega Knowledge Prompt Manager and creating prompts in Pega Knowledge is located here: Pegasystems Documentation

Sample Pega Knowledge Prompts are located in this article: Knowledge buddy prompt instructions - #3 by grafg - General - Pega Forums

Thanks for the “how to” on this, @grafg

Thank you for documenting this configuration and for highlighting how Assisted authoring with Pega GenAI fits into Pega Knowledge.

This is a good example of GenAI being applied in a constrained, author facing context where consistency and governance matter more than open ended creativity.

The fact that prompts are defined and managed centrally by a prompt manager, and that authors invoke them through controlled options, reinforces Pega’s approach of design time governance rather than ad hoc generation.

It is also useful that you call out the dependency on licensing and cloud deployment, as those are often overlooked when teams first explore these features. From a broader perspective, this pattern shows how GenAI can improve productivity without becoming the source of truth itself.

Knowledge quality, approval workflows, and ownership remain unchanged.

I would be curious to hear from others how they have designed prompt governance for Knowledge authoring, and what review or validation steps they have found necessary to maintain trust in AI assisted content over time?