I am raising this enhancement request on behalf of Rodrigo from Feedzai, focusing on improving how AI generated answers are managed within the Content Library. Today, when AI Draft pulls information from the Content Library and generates a response, that output is later saved as a new question answer pair even when a contextually similar QA already exists. Because AI responses vary slightly in wording, these entries are treated as unique, which leads to rapid duplication and Content Library sprawl, and the customer does not intend to enable verbatim since like to like responses are not the goal here. While features such as content cleaner and view similar content exist and enable the client to merge duplicates, they rely on string based matching and require manual one by one review, which does not scale for customers generating content at volume through AI Draft. Bulk merge also does not meet the customer request, as in scenarios where a project has thousands of QA pairs, those still need to be reviewed and merged individually rather than in a truly grouped or contextual manner, which is what the customer is requesting. The customer is looking for a smarter capability where the system can contextually identify similar questions and answers even if phrasing differs and proactively recommend grouping or consolidating them, such as indicating that similar QAs already exist and allowing users to group them under a primary question or treat others as alternates in bulk. The goal is to reduce duplication, improve Content Library governance, lower manual effort, and reduce the risk of outdated or conflicting answers. This request spans both AI and Content Library domains and would require collaboration across teams as it impacts AI content generation and downstream Content Library management.
Special note from the Customer: "I see this as a change that could significantly transform the Responsive product and have a strong customer impact, as it would drastically reduce the resources needed to manage the Content Library."
Looking forward to your thoughts and would be great if this could be prioritized.
CC: @sathish s @Rithvik Sridhar
