Version 1.2

Conversational AI 2026.1.0

Hamlet search results display capability in Knowledge Discovery for End Users

The Knowledge Discovery for End Users widget now features the Hamlet search results display capability. End users can now see search results from the Hamlet engine directly in the conversational interface, improving answer discoverability

Improvements

• Users can now use image attachments in conversation contexts during userWidget interactions. This functionality enables vision-based AI capabilities and supports multimodal conversation flows under Vision development.
• Improved the formatting for chatbot conversation history in PDF email transcripts when RAG (Retrieval-Augmented Generation) answers are present. RAG-generated responses now display with proper formatting and structure, making email transcripts more readable for end users.
• Added self-service provisioning and configuration capabilities for TeamsBot directly into the CAI Admin Panel. Administrators can now deploy and configure MS Teams integrations without manual backend intervention, streamlining tenant onboarding.

Bugs Fixed

• Fixed a bug that caused the user widget (userWidget) to behave incorrectly in certain edge cases in the Safari browser.
• Fixed scroll-to-bottom button behavior in userWidget. The button now correctly hides when the conversation is already scrolled to the bottom, eliminating visual inconsistency and improving UX.
• Fixed a userWidget height recalculation issue on mobile devices after closing the keyboard. The widget now correctly adjusts its viewport height, preventing layout overflow and improving mobile UX.

AI Core 1.2

AI Core 1.2 expands the platform from a RAG-focused service into a broader agentic AI platform. The main addition is the new Agents subsystem: tool-using AI assistants with chat, scheduled and agent-to-agent triggers, long-context support, dedicated metrics, source and tool filtering, and admin UI support.

The release also strengthens governance, observability and operational control. Content-safety checks now cover more interaction points, access-control support has been extended for Confluence and Matrix ESM, citation extraction improves source attribution, and new telemetry gives teams a clearer view of LLM, embedding, tool and agent activity.

The tools ecosystem has also been modernized around JavaScript skills, MCP servers and built-in platform tools. AI Core 1.2 introduces a dynamic Tools Repository UI, updated MCP integration, crawler-only mode for externally orchestrated ingestion, broader automatic testing, UI/UX improvements, dependency updates and bundled stability fixes across the stack.

Key Features

Agents and Multi-Agent Workflows

• New Agents capability for building tool-using AI assistants, with chat support, streaming responses, sessions, rate limits and platform controls.
• Support for chat, scheduled and agent-to-agent triggers, including overlap protection, graceful shutdown and internal-service authentication.
• Agents can now call other agents, with safeguards to prevent runaway workflows.
  Built-in agent tools for Agentic RAG and image analysis.
• Configurable source and tool filtering per agent.
• Agent support for client-provided message history, extended response metadata and context-status reporting.
• Context summarization for longer conversations, so older turns can be summarized instead of simply removed.
• Dedicated agent logs, trigger analytics and token-usage views.

Governance, Access Control and Content Safety

• Content-safety checks expanded to images, tool outputs and ingestion workflows.
• Configurable regex-based pre-filters for fast handling of known unsafe or malformed prompt patterns.
• Confluence permissions are now respected more consistently, including safer behavior when user context is incomplete.
• Matrix ESM access-control handling improved for cases where end-user context is unavailable.
  Secret handling has been standardized, improving consistent redaction and preservation of sensitive fields.
• Additional hardening around logging, configuration constraints, authentication fallback and admin actions.

RAG, Knowledge Quality and Attribution

• Document quality checks during indexing, including text-length, word-count, character-ratio and LLM-based assessment rules.
• Structured quality-check results surfaced in the admin UI with progress reporting.
• Citation extraction to identify which retrieved documents were actually used in a response, improving verifiable source attribution.
• RAG Index upload flow expanded for multi-file and folder uploads, with managed upload identifiers instead of exposed server paths.
• Improvements to source matching, result ordering, full-text search synchronization and document filtering.

Tools and Integrations

• New Tools Repository for managing tools more easily, with editing, validation and hot reload support.
• Tools are now grouped into JavaScript skills, MCP servers and built-in platform tools.
• MCP integration has been updated with improved streaming, connection handling and session support.
• Independent enablement controls for Passthrough and RAG usage across tool types.
• Tool access can be configured per agent, including which tools are available and how long they may run.
• Matrix42 tool coverage has been expanded for announcements, orders and ticket workflows.
• Improved end-user authentication handling for Matrix42 tools.

Telemetry and Analytics

• New LLM telemetry capturing provider, model, task type, token usage, duration, errors, streaming mode and transaction IDs.
• Embedding telemetry with daily aggregation by date, model and backend.
• Better tracing for LLM API calls.
• Token analytics split by input/output and task type, with dedicated Tools and Embeddings views.
• Tool success/error analytics and date-range filtering for security analytics.
• Reusable feedback/rating system across RAG, Passthrough and Agent chat.
• API cleanup and improved security-statistics filtering.

Automatic Testing

• New Automatic Tests module for RAG quality evaluation, multi-turn conversations, CSV import/export and dedicated test storage.
• Additional test suites for embeddings, firewall, full-text search and pre-LLM behavior.
• Agent workflow tests with tool-trace capture and LLM-based evaluation.
• Broader unit and integration coverage across the admin UI, API, crypto, database tools, metrics, tools and user-documentation modules.

Crawlers and Operational Modes

• New crawler-only mode exposing a stateless REST API for Confluence, Web and SharePoint crawls.
• Direct write support to Azure Blob Storage for externally orchestrated ingestion scenarios.
• Crawler and plugin stability improvements, including progress reporting, HTTP error handling, pagination handling and CSV edge cases.

Admin UI and UX

• Unified admin layout with shared shell/sidebar/main patterns.
• Improved responsive and mobile experience.
• Better agent chat UX, expandable rows, source lists, scroll behavior and chat-box sizing.
• Refined logs, search, ratings and analytics views.
• Improved configuration validation, field bounds, tab highlighting, notifications and documentation links.
• Updated architecture diagrams and agent documentation.

Improvements and Maintenance

• Agent configuration and lifecycle refinements, including UUID-based identifiers, improved context behavior, cleaner tool-reference handling and updated docs.
• RAG and search refinements, including BM25/full-text search behavior, source ordering, filtering and indexing synchronization.
• Tools and MCP cleanup, including modal validation, clearer tool names, structured logging and cleanup for legacy tools.
• Platform maintenance, including dependency updates, removal of deprecated imports, Ubuntu-based Dockerfile support, tini init process, Chromium process handling and Consul-token path updates.
• Admin UI cleanup, including reduced duplicated log/search code, better sidebar reuse, expanded documentation links and refined test-tab workflows.

Bundled Fixes

Agents and Tool Execution

• Improved usage accounting, tool/source access checks, full-text search source handling, rate-limiter behavior and save-failure handling.
• Fixed context propagation and several edge cases in agent tool execution.

RAG, Search and Embeddings

• Corrected source ordering, source matching, document ordering and BM25/full-text search behavior.
• Improved embedding token accounting, URL matching in tests and handling around zero-vector normalization.
• Removed an unnecessary dependency from the search stack.

Crawlers, Plugins and Ingestion

• Improved crawler behavior around HTTP errors, pagination, CSV trailing rows and progress reporting.
• Added resilience improvements for plugin failure modes and ingestion workflows.

Tools and MCP

• Improved passthrough tool identity propagation.
• Addressed MCP test-connection and long-lived stream handling.
• Corrected built-in tool classification and related metrics categorization.
• Consolidated smaller tools-package fixes.

Admin UI and API

• Fixed tab highlighting, test-tab flows, long-sidebar chat sizing, modal submit behavior, clipboard handling and loading-state edge cases.
• Improved Azure OpenAI deployment-name handling for names containing dots.
• Fixed admin UI tests affected by agent-list behavior.

Access Control, Secrets and Platform Stability

• Strengthened access-control handling for requests without complete end-user context.
• Improved secret-field redaction and preservation when editing sources.
• Added process-management improvements for Chromium-based workloads.
• Bundled minor typo, cleanup and small-fix changes.

Version 1.1.1

Knowledge Discover Package Installer

Bug Fixes

• Fixed an issue that prevented successful environment updates by not removing certain obsolete artifacts. These leftover files required manual intervention and could block update procedures.

Conversational AI 25.3.2.2

No changes to Conversational AI for this release.

AI Core 1.1.22

New Features

LLM Support

• Added support for Mistral reasoning models.

Version 1.1.0

Conversational AI 25.3.2.2

Improvements

Agent Widget

• Improved session handling with partitioned cookies: Introduced partitioned cookies to enhance session management. This improves session stability, prevents cross-site tracking issues, particularly in browsers with strict cookie policies (e.g., Chrome's third-party cookie restrictions).

KD package Installer

• Added chatbot database initialization progress logs: Introduced detailed logging during chatbot database initialization in the Knowledge Discovery installer. Logs now display progress stages, timestamps, and status messages, improving visibility and troubleshooting capabilities for on-prem deployments in air-gapped environments.

Internal Improvements

• Code quality and stability enhancements across chatbot and voicebot components.

Bug Fixes

Chatbot Widget

• Fixed widget size reset on scroll (desktop): The chatbot widget no longer unexpectedly resizes when users scroll the page on desktop browsers. This ensures a consistent, stable widget experience during interactions.
• Fixed input text color in dark mode: Corrected text color rendering in the chatbot input field when dark mode is enabled, improving readability and visual consistency.
• Fixed link clickability issue: Resolved a bug where links in chatbot messages were not clickable due to the "scroll to bottom" button overlaying message content. Users can now interact with all links in chat history without obstruction.
  Agent Widget

Admin Panel

• Fixed GenAI response statistics display: Corrected the display of statistics related to generative AI responses in the Reports / Chatbot / Usage section. Metrics now accurately reflect GenAI usage, enabling better monitoring and analysis of AI-driven interactions.

AI Core 1.1.21

New Features

Integration Capabilities

• Microsoft SharePoint Connector: Support for Action Credits consumption model
• AWS S3 Source Support: Integration with S3-compatible storage solutions for document retrieval
• Apache Shibboleth SSO: Single sign-on integration for enterprise authentication

M42 Platform Integration

• M42 Nest Billing Integration: Support for Action Credits consumption model
• M42 Enterprise RBAC: Document scope limitation based on user permission profiles
• M42 Enterprise CI Support: API integration for Tickets, Announcements, and Services
• M42 Pro REST API: Native REST API support within JavaScript tools with automated discovery

Document Processing

• Structured Data Extraction: Advanced data extraction with query capabilities optimized for range-based operations where SQL outperforms embeddings or full-text search
• Structure-Aware PDF Splitting: Intelligent PDF document segmentation based on document structure
• LLM-Based Chunking: Semantic document chunking based on contextual meaning rather than structural boundaries or character count
• Automated Link Downloader: Dynamic website content fetching to provide contextual awareness of the user's current browsing session

LLM Support

• Mistral.AI Integration: Added support for Mistral.AI language models

Improvements

RAG and Content Management

• Tool Integration in RAG Mode: The model now intelligently determines whether to utilize RAG for answering queries or invoke tools for data access, including REST API consumption
• Multi-Database RAG Support: Introduced the ability to maintain separate RAG sources, enabling segmentation by audience or use case
• Cross-Engine Context Preservation: Implemented seamless conversation continuity when transitioning between different chat engines
• Improved Query Condensing: Enhanced handling of extended conversations that exceed LLM context limits, with optimized RAG document selection throughout the conversation lifecycle

Security and Administration

• Secrets Management for JavaScript Engine: Implemented secure environment variable access through getters in JavaScript sandbox, preventing exposure of sensitive credentials in source code
• Automated Log Retention Policies: Added configurable chat transcript retention with automated conversation removal capabilities
• Knowledge Base Administrator Role: Introduced new limited-privilege administrative role for UI management
• Rate Limiting and IP Whitelisting: Enhanced network security controls for access management
• Comprehensive Audit Logging: Detailed tracking of administrative actions within the UI

Performance and Accuracy

• Reranking Optimization: Refactored the reranking logic for improved RAG performance
• Token Counting Precision: Enhanced accuracy in calculating document and message lengths for context window management
• Extended Metrics API: Expanded API endpoints for system metrics and monitoring

User Interface

• Logs and Audit Section: New dedicated UI section for log review and audit trail examination

Bug Fixes

• Resolved performance and search issues in log browser functionality
• Implemented a fallback mechanism for missing session headers from client requests
• Corrected keyword extraction in full-text search for specific model configurations
• Fixed parsing errors with certain DOCX formatting specifications
• Prevented log overflow when recording tool execution traces
• Resolved silent failure scenarios in Distiller when the LLM API becomes unresponsive

Version 1.0.20250925.1 

The 1.0.20250925.1 release is the initial version of the Knowledge Discovery for End Users extension.