This update introduces behavioral intelligence capabilities to the Virtual Assistant and expands Knowledge Base management options with automated synchronization and content segmentation.

New Features

Contextual Intelligence (User Archetypes) The assistant now utilizes an internal profiling system to adapt response output based on the user's engagement level and tenure.

• Behavioral Adaptation: The system identifies the user's current career stage (e.g., Newcomer, High Performer, At-Risk) to tailor recommendations.
• Dynamic Response Logic: The LLM adjusts its tone and suggestion types accordingly, prioritizing instructional guidance for new users versus retention and wellness strategies for disengaged users.

Automated Knowledge Synchronization Added a new method for populating the Knowledge Base. The system now performs a daily crawl of all content published on the platform's public pages, automatically indexing and cataloging this information for immediate retrieval by the Bot.

K-Anonymity Privacy Protocol Implemented K-Anonymity standards for the new profiling features. This ensures that individual user archetypes are not identifiable in administrative reports unless a minimum group size threshold is met, preserving individual privacy while allowing for personalized assistance.

Improvements

Knowledge Base Segmentation Enhanced the manual PDF upload functionality. Administrators can now assign visibility rules to specific documents or categories. This allows content to be restricted to specific user groups or roles (e.g., restricting management policies to the "Managers" group), ensuring that the Bot only retrieves information the user is authorized to view.

Query Routing Optimization Updated the intent classification logic to improve how the Bot routes questions to specific domains (HR, IT, Organization), resulting in faster and more accurate retrieval of information.