📈Features & Updates
This section provides release notes and the latest product updates from Blockbrain. Users can learn about new features, improvements, and fixes, ensuring they fully leverage the tool's capabilities.
Product Updates - August 26th, 2025
New LLM: Gemini 2.5 Flash Lite - New Gemini 2.5 Flash Lite provides quicker responses while maintaining quality.
Automatic Model Selection (GPT-5 Auto) - Automatic model choosing intelligently picks the best model out of the GPT 4 or GPT 5 family for your question.
Cleaner Chat Interface - Redesigned layout with organized Options menu in the sendbox (with a new “Writing Style” option) and streamlined file management.
New LLM Selection Layout - Improved positioning moves AI model selection and token usage to the top-left corner.
Enhanced File Management - Streamlined data room experience with search and bulk selection capabilities. Use the search bar and checkbox selection in your data room list.
Improved Settings Panel - Improved settings panel with better organization and navigation. Access through the settings icon with new close buttons and enlarged clickable areas.
Dynamic Insight: Flexible Web Search - Flexible web search with different search types in Dynamic Insight (Web Search or Web Search Pro).
Full Conversation Downloads - Full conversation downloads including entire chat history and sources. Use the Export PDF button next to a data room.
Improved Web Crawling - Improved website crawling for more reliable information extraction — now applied automatically when crawling a website in a database.
Upload Files in Chat Interface - Upload files in chat via the file upload button directly in the sendbox or drag & drop into the window.
ZIP File Upload - ZIP file support for uploading multiple documents at once. Upload ZIP files through the standard file upload interface.
Default LLM Selection (Editors only) - Bot customization with specific AI models set as defaults. Administrators can configure default LLMs in bot settings.
Updated Embedding Model Descriptions - Updated model information with clearer embedding model descriptions. View improved descriptions in the model selection interface.
British and American English Option - Regional language support with British and American English variants. Select your preferred language variant in platform settings
Workflow Auto-Advance Toggle Behavior Improvements - Auto-advance toggle refinements with better behavior in workflow modes. Configure auto-advance settings in workflow configuration.
General UI Enhancements - General UI enhancements with visual improvements across the platform. These improvements are visible throughout the updated interface.
1. New LLM: Gemini 2.5 Flash Lite
A lightweight AI model optimized for speed that provides faster responses for routine tasks and quick interactions.
*New LLMs must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
2. Automatic Model Selection (GPT-5 Auto)
An intelligent system that automatically selects the most appropriate GPT 4.1 or GPT 5 model (8 different LLMs) based on your query's complexity and requirements. Select "GPT 5 (Auto)" from the LLM-model dropdown to enable automatic routing, and you'll see which specific model was chosen for each response.
*New LLMs must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
Image 1: Select GPT 5 (Auto)
3. Cleaner Chat Interface
A comprehensive interface redesign featuring a new Tools menu that consolidates Writing Styles and Prompt Library (prompts can be adjusted through the Sendbox), plus enhanced data room management with search and sorting. Access the tools menu from the main chat interface and use the new search functionality in your data room list.
Image 1: Upload your files
Image 2: Select your prompts or the bot’s writing style
Image 3: Select your writing style
Writing styles essentially control how much the AI elaborates on its answers, ranging from very short answers (1 or 2 sentences) to comprehensive deep-dives, allowing you to match the response length to your specific needs:
Auto - Automatically determines the optimal response length based on your question
Very Short - Minimal, concise answers with essential information only
Short - Brief responses that cover key points without elaboration
Normal - Standard response length with balanced detail and context
Long - Extended responses with comprehensive explanations and examples
Extra Long - Maximum detail with thorough analysis and extensive context
Image 4: Web Search Web Search: Perfect for quick facts and straightforward answers. This mode executes 3-8 targeted searches simultaneously to deliver instant, relevant results. Ideal for definitions, current events and simple requests.
Web Search Pro: For complex investigations, the systems splits your query into 3 core themes, runs 9-24 searches (3-8 per theme) and delivers structured, multi-perspective results, which are ideal for market research, academic topics and strategy.
4. New LLM Selection Layout
LLM selection dropdown and token usage display have been moved to the top-left corner for better visibility and easier access. Look for your model selection and token counter in the new prominent top-left location of the chat interface.
Image 1: New LLM selection in the top left of the data room
5. Enhanced File Management
Enhanced data room interface with compact displays, search capabilities, and multi-selection to e.g. delete data rooms. In order to use the bulk selection, go to the left sidebar and select the dots-icon next to the magnifying glass and click “select”.
Image 1: Select the dots-icon
Image 2: Click “Select” to access the bulk-selection or sort your data rooms with the options above
6. Improved Settings Panel
Redesigned settings interface with close buttons, better spacing, and enlarged clickable areas for improved navigation. Access settings through the settings icon and notice the new close buttons and improved layout throughout the interface.
7. Dynamic Insight: Flexible Web Search
Enhanced web search in Dynamic Insight* with selectable search types for more targeted results. Choose your preferred search type from the new dropdown menu when configuring web search in Dynamic Insight.
Image 1: Select your Web Search type
*Dynamic Insight is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
8. Full Conversation Downloads
You can now not only download messages, but entire data rooms as PDF, by clicking on the “dots” icon and selecting “Export PDF” on the left side of the data room.
Image 1: As a new feature you can now download you entire data room as PDF
Feature tip: Include entire chat history in the insight
When creating an insight from a chat message, you can enable a simple toggle (“Include entire chat history in the insight”) to save the entire chat history in the Insight.
When this Insight is linked to a data room, the entire chat history is taken into context—not just the last bot response.
9. Improved Web Crawling
Upgraded web crawling* system using Firecrawl technology for more reliable information extraction from websites and documents. This enhancement works automatically in the background when processing web content for your knowledge bases.
*only visible for Builder and Admins
10. Upload Files in Chat Interface
Add files directly within your conversation without switching screens or interrupting your workflow. Use the file upload button located directly in the chat sendbox or drag and drop files into the screen for immediate processing.
Image 1: Either click “Upload Files/Folder” or drag and drop your files into the data room
Image 2: Your file(s) will be shown to you in the sendbox and are automatically connected with the data room through “Files”
11. ZIP File Upload
Upload and process ZIP files containing multiple documents simultaneously, with automatic extraction and processing of all contained files. Upload ZIP files through the standard file upload interface or via drag and drop, and the system will automatically extract and process all contents.
12. Default LLM Selection (Editors only)
Editors can now set specific LLMs as favorites for individual bots based on their intended purpose and use case. Navigate to bot settings as an Editor and configure default LLMs in the bot configuration section.
Image 1: In the top left of your data room, click the selected LLM you chose
13. Updated Embedding Model Descriptions
Clearer and more detailed descriptions for embedding models to help with informed selection decisions. View the improved descriptions when selecting embedding models in the admin dashboard under the "Embedding Models" section.
14. British and American English Option
Added regional language variants with British and American English bot-options for improved localization. Select your preferred language variant in the data room settings under language.
Image 1: Select your preferred bot language
15. Workflow Auto-Advance Toggle Behavior Improvements
The Auto-Advance toggle now works consistently across workflow modes. In Human-in-the-Loop mode, ON automatically progresses to the next step while OFF requires manual advancement. In Autopilot mode, the toggle has no effect and maintains automatic flow.
Image 1: Enable the “Auto-Advance” button to automatically progress to the next step
16. General UI Enhancements
Platform-wide visual improvements including layout optimizations, visual consistency updates, and overall user experience enhancements. These improvements are visible throughout the updated interface and require no user action to access.
Product Update – August 11th, 2025
1. New Models
Integration of new LLMs*
OpenAI
GPT-5 Family – Next-gen flagship with major performance boosts across 4 modes (hosted on EU & US servers)
GPT-5 – Specialized conversational model with superior dialogue skills (likely to replace GPT-4.1 as the default Company GPT LLM)
GPT-5 Thinking – Flagship variant with advanced reasoning power
GPT-5 Mini – Lightweight, speed-optimized version
GPT-5 Nano – Ultra-fast model for real-time use
GPT-image-1 – Image Generation is now available on Azure
GPT 4o Realtime audio – Realtime Chat via Voice Mode
All LLMs are hosted on servers in the EU (and optionally in the US).
Anthropic
Claude 4.1 Opus – Most powerful model from Anthropic with advanced reasoning capabilities
Hosted on US servers; EU availability coming soon.
Meta
Llama 4 is Meta’s latest generation of AI models, offering advanced reasoning, long-context handling, and powerful multimodal capabilities.
Llama 4 Maverick – High-performance general-purpose model excelling in reasoning, coding, and multilingual capabilities.
Llama 4 Scout – Optimized for ultra-long context and multimodal tasks, ideal for document and image analysis.
Hosted on US servers; EU availability coming soon.
Mistral
Mistral offers three different models, depending on complexity.
Mistral Small: Fast and efficient, best for simple tasks.
Mistral Medium: Balanced performance, suitable for moderate complexity tasks.
Mistral Large: High performance, ideal for complex and detailed tasks.
Hosted by Mistral in the EU.
*New LLMs must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
Enhanced Transcription Capabilities*
Realtime Transcription Model – Seamless voice-to-text conversion with improved accuracy and faster processing
*This feature is still in early alpha, and its quality doesn't yet match the current version. It is controlled via a feature flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
2. Enhanced User Experience
Improved Content Management
Folder Upload Functionality – Upload entire folders to data rooms with a single action, making bulk content management significantly faster
Preserved Formatting – Blank lines in bot "Initial Instructions"* are now maintained after saving, ensuring your formatting stays exactly as intended
*Initial instructions in bots are only visible for Editors.
Enhanced Chat Experience
Unified Chat History – In Webcomponents (integration in SharePoint, Microsoft Teams or a website) private login users can now view consolidated chat histories across all sessions for better continuity
Mobile-Optimized Interface – Web app adjustments provide enhanced usability and navigation on phone devices
Voice Mode: Realtime Chat (Alpha)*
Voice Mode Integration – Experience low-latency, "speech in, speech out" conversational interactions with enhanced responsiveness via Azure GPT 4o Realtime audio for a more natural conversational flow
Simple Activation – Simply activate "Voice Mode" next to the sendbox to start talking with the AI
Note: This feature is currently in alpha and may still contain bugs
*Realtime chat is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
3. Collaboration & Management Enhancements
Advanced Role Management
Enhanced Consumer Permissions – Consumers can now be assigned as bot editors, allowing them to modify settings of existing bots (note: editors cannot create new bots)
Improved Editor Workflow – The “Remove Editor” option has been replaced with a new Demote to "User” option.
*Sharing Settings in bots are only visible for Editors.
Administrative Controls
Image Generation Settings – You can control image generation capabilities on a per-bot basis through dedicated bot settings*
*Image generation is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain). The Bot Settings are only visible for Editors.
Data Retention Enhancement – Comprehensive data retention management with informational layers and extend retention for better compliance control
Improved Insight Management
Streamlined Insight Sharing – Enhanced sharing workflow with improved user experience
Smart Error Handling – "Insight Unavailable" modal for deleted insights provides clear communication and next steps
4. UI/UX Improvements
Enhanced LLM Selection Interface
Performance Visualization – Updated LLM list view with clear performance KPIs for informed model selection
Improved Visual Navigation
Bot Status Indicators – Clear status icons for scheduled-deletion and template-based bots in dropdown menus
Enhanced Sorting – Organized bot dropdown in analysis tab for faster navigation*
Updated Tooltips – Clearer descriptions for right sidebar buttons in Knowledge Bot chat interface
*Analysis is only visible for Admins.
Streamlined User Interface*
Refined Connection Modals – Updated "Disconnect" modals for Azure Group and SharePoint with improved clarity
*This is only visible for Admins.
5. Integration Enhancements
SharePoint Integration
Global Floating Button – Webcomponent integration as a floating button across SharePoint environments
Configurable Sidebar – Control sidebar state during webcomponent initialization for customized user experiences
Administrative Enhancements*
Organization ID Display – The Org ID is now visible in the frontend for easier identification and support. Simply click your profile and copy the Org ID displayed beneath your name and email.
*This is only visible for Admins.
Product Update – July 28th, 2025
1. New Models
Integration of new LLMs*
o3-pro – hosted on EU & US servers
Gemini 2.5 Pro (Google) – hosted on EU & US servers
Gemini 2.5 Flash (Google) – hosted on EU & US servers
Grok 4 (xAI) – hosted on US servers; EU availability coming soon
Integration of new Embedding Models*
Gemini Embedding 001 - hosted via Google on US-servers (EU availability coming soon)
*New LLMs and Embedding Models must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
2. Enhanced User Experience
Streamlined Content Management
Improved Editor - Edit bot responses (just double click on an answer) and insights with fewer clicks.
Access formatting tools more easily, and enjoy a more intuitive workflow that makes content updates significantly faster.
Listen to Bot Messages
Listen to bot responses read aloud by clicking “Listen” (headphone icon).
Control playback speed or pause using the media panel while continuing to navigate the platform. You can even download the bot response read aloud.
3. Collaboration & Managemenet Enhancements
Bot Creator Update: Streamlined Creation with Nexus by Default
Faster Workflow: Removed the “Pick a Bot Template” step for quicker setup
Simplified Experience: Nexus is now the default and only option
Instant Start: Go directly into building your bot without extra clicks
*only visible for Builder and Admins
Default Web Research*
New “Web Research” category in bot settings allows enabling or disabling Web Research per bot
Option to enable Web Research by default automatically activates Web Research for all newly created Data Rooms, eliminating the need for manual activation each time
Set default Web Research type: Web Research, Web Research Pro, or Web Research Pro (R)
*only visible for Builder and Admins.
Role Management
Direct Role Upgrades - Upgrade user roles directly from the dropdown menu in bot Share settings from User to Editor*
Peer-Level Role Management in Knowledge Databases - Content Managers and Contributors can now promote, demote, or remove users at or below their own level
*only visible for Builder and Admins
4. Content Organization
Include entire chat history in the insight
Enable a simple toggle (“Include entire chat history in the insight”) to save full chat histories in the Insight.
When this Insight is linked to a data room, the entire chat history is taken into context—not just the last bot response.
Improved Bot Sharing Visibility
Bots now clearly display all sharing states (private, restricted, public) in both the bot list and bot creator tab
5. UI/UX Improvements
Agents are now Prompts / Prompt Library
We’ve updated the terminology to make the feature’s purpose clearer and more intuitive.
Functionality remains unchanged—only the name and design has been updated.
New LLM Selection Design
We’ve redesigned the LLM selection to make it clearer and faster to use.
Search Function: You can now quickly find and select the LLM you want from all models available on your platform.
Information Button (ℹ): Click the info icon to view detailed information about each LLM.
Key Metrics: The evaluation criteria have been refined for better clarity. Models are now rated by Answer Quality, Speed, and Cost Efficiency.
Multiselect for E-Mails
Select all E-Mails with a single click
Speech-To-Text: Intuitive Button Placement
Enhanced Layout: Speech-to-Text button repositioned within the "Send-Box" for greater prominence
Improved Discoverability: More intuitive placement based on user expectations and the upcoming Voice Mode (real time chat with the bot)
Branded App Icon / Favicon in the Browser Tab
Professional appearance in the browser tab and when added to home screens on iOS, Android, or Windows
Consistent Terminology Update across the platform
“Databases” is now called “Database Sources”; the word “connected” has been removed from the Knowledge Management section of data rooms.
“Knowledge Destinations” is now called “Destination Databases”.
Improved OCR Processing
Better handling of two-column PDF layouts for more accurate text extraction
Product Update – July 15th, 2025
Intuitive Drag & Drop Functionality
Easily move files and Insights into folders inside the platform with a simple drag & drop action
Visual indicators highlight valid drop targets when dragging files
Receive clear visual confirmation when files are successfully moved
Get helpful feedback when attempting invalid file moves
Automatic Removal of Disconnected Databases from Data Rooms
When a database is disconnected from a bot, it will automatically be removed from existing data rooms.
Receive clear alerts when a bot is no longer connected to a database — either upon entering a data room or during interactions.
Alerts remain visible until dismissed, ensuring awareness of potential impacts on response quality.
1. Enhanced Analytics & Exports
Improved Usage Analytics*
Track real-time token and storage usage through comprehensive dashboards.
Monitor current consumption, projected limits, and per-bot statistics for better tracking of usage per use case.
*only visible for Admins
Excel Table Export
Instantly export tables from prompt results to Excel (.xlsx) with one click, maintaining original formatting and structure for easy analysis and sharing.
Only table data is exported; non-table text is excluded.
If no table is detected, the “Download as XLSX” button is disabled.
Image Upload Token Tracking
Token consumption for image uploads via Prompt Injection is now tracked and displayed in the usage dashboard.
Provides transparent resource cost monitoring for both individual requests and cumulative usage.
2. Knowledge Management Improvements
Sorting Insights by Creation
Sort insights by creation date with the new "Created At" column in table view, making it easier to locate original entries even after batch updates.
The less relevant node type column is automatically hidden when space is limited.
Folder Descriptions
Add descriptions to Insight folders to better organize and understand folder contents.
Descriptions appear as sublines in the folder list, can be entered via text or voice, and are automatically shortened with full text available on hover.
Folder Selection in Insight Editor
Choose or create target folders directly when creating or editing insights.
Your content is automatically saved or moved to the selected location, eliminating extra organizational steps.
3. Interface Refinements
Improved Bot List View
The bot list (former dashboard) now features a lighter card design for better visibility.
It defaults to a two-column layout for easier browsing but remembers your preference if you want to keep the one-column by default.
Better Deleted Resource Handling
Deleted bots or data rooms now display their last known name in grey with a "deleted" icon in workflows and results.
This preserves context and avoids confusion when reviewing past activities.
Simplified Data Room Creation
Data rooms now receive automatic names by default, eliminating the "Name your Data Room" prompt during creation.
You can still rename data rooms later if needed.
Enhanced Text Editor
The editor toolbar now includes all essential markdown formatting options: headers, bold, italic, strikethrough, code, lists, quotes, alignment, hyperlinks, and tables.
4. New Large Language Models
Integration of new LLMs*
Llama models (from Meta) now available on US-servers (EU availability coming soon).
Gemini 2.5 Pro now available on EU Servers.
*New LLMs must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
Product Update – July 1st, 2025
1. Organization & Management
Automatic Naming of Data Rooms
System now automatically names new data rooms based on your first message
Helps organize your data room list more efficiently
Prevents multiple generic "New Data Room" entries
Automatic naming occurs only once and won't override custom names you set
Folder Selection in Insight Editor
Choose where to save insights during creation or editing
Current folder path displayed below insight title
Browse existing folders or create new ones via dialog
Insights automatically stored in chosen folder for better organization
2. Interface Enhancements
Redesigned Bot List Page (former “Dashboard”)
Clearer, more intuitive bot overview interface
Improved navigation and organization with more filtering options
Display of Complete File Path for References
Complete file path now visible for all database source references
Provides enhanced context for source identification
Improves traceability of information across the platform
Accessible in all reference displays throughout the system
Newsletter Language Preference*
Go to your profile (top right corner) and click “Manage Subscriptions”
Select your preferred language for newsletters related to knowledge contributions to a knowledge base
All future emails will be sent in the language you selected
*The newsletter is part of the ‘Knowledge Contribution Subscription’ and controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
3. Workflow Improvements*
Step Result Binding
Choose whether workflow steps use previous step results as input
Visual diagram with connection points shows data flow between steps
System prevents illogical connections between steps
Default setting requires explicit enabling of connections
Multi Data Room Selection in Workflows
Select specific data rooms as input for each workflow step
Toggle data room selection with administrator controls
Choose multiple data rooms from searchable list with checkboxes
Selected data rooms appear as removable tags for easy management
*Workflow is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
4. Enhanced Bot Access Control
Update Access Levels for Connected Database Sourcess in Bot-Sharing
Control bot access with Private, Restricted, or Public sharing modes
Connected database sources automatically update access settings to match bot privacy
Clear notifications guide you through privacy changes
Ensures data security and transparency about bot access
Bot Status Update on Privacy Change
Shared bots made private appear as "offline" in your bot list
View old conversations while new sessions are restricted
System provides popup and permanent notifications until acknowledged
5. Mobile Experience*
UI Improvements in Chat View
Enhanced chat interface for smoother mobile experience
More intuitive navigation and interaction
Add Photo Functionality to Chat View
Send photos directly from the mobile app chat view
Seamless media sharing capabilities
*The mobile app must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
6. Integration
Web Component of Chat in MS Teams*
Blockbrain chat now embeddable directly in Microsoft Teams
Seamless integration with existing interface
*If you are interested in the chat integration with Microsoft Teams, please contact the Key Account Manager at Blockbrain.
Integration of OpenAI o3-pro Model
Now available on US-servers (EU availability coming soon)
Designed for advanced reasoning and generation tasks
Folder Navigation Improvements
Clicking anywhere on a folder now opens it directly, showing its contents
Consistent behavior implemented across file browser, database source, and SharePoint interfaces
Eliminates the need to click on specific small areas for navigation
Provides a more intuitive experience aligned with standard file system behaviors
Expandable Prompt Input Fields
All prompt input areas now feature expandable text boxes
Easily compose and review longer, complex prompts with improved visibility
Consistent with the existing functionality in the Workflows section
Enhanced user experience when working with detailed prompts
Standardized Sharing Settings Interface*
Consistent "Sharing Settings" tab format throughout the platform
Same interface during both bot creation and subsequent editing
Eliminated confusing toggle switch in favor of a unified approach
Ensures seamless transition between creation and management workflows
*only visible for Builder and Admins
Multi-Select Connect/Disconnect for Files
Select multiple files at once for bulk connection/disconnection in a dataroom
Significantly improves efficiency when managing large numbers of files
Streamlines prompt setup and modification workflows
Web Research Enhancements*
New “Web Research” category in bot settings to enable or disable Web Research per bot
Option to enable Web Research by default for new Data Rooms
Removes the need for manual activation in every new Data Room
*only visible for Builder and Admins. This is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
Enhanced Search Capabilities
Global search now includes Insight folders
Folders appear with icons and full path information
Direct navigation to folder contents with a single click
Optimized for quick performance with thousands of folders
Document Processing Enhancements
New option to combine multiple PDF pages into a single chunk (this has to be activated in the "Advanced Settings" of a Knowledge Base)
Support for IFC file format via Knowledge Base integration
LLM Gateway integration for improved reliability and load balancing
Enhanced processing of images and tables within documents
Interface Refinements
Added separator line in "Save Insight" modal for improved visual clarity
Fixed "Save in Database Source" modal for workflow outputs*
Added "(Deleted)" placeholder for removed Bots/Data Rooms in workflow records*
Split Azure App Registrations for smoother Microsoft service connections
Automatic full-screen view for long chat entries
*Workflow is controlled via a Feature Flag (Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
Product Update – June 10th, 2025
New Bot-view in Main Navigation
“Bots” receives a dropdown menu
Recently used Bots now appear directly in the main navigation
Enhanced with search functionality for quicker access
Use "Show Bot List" to see all available Bots
Workshop is now called Bot Creator (only visible for Admins and Builder)
Data Retention Option for Data Rooms
It is now possible to auto-delete data rooms after a predefined time
Configure data retention policies directly in Bot Settings under "Data Retention" (Bot Editors only)
Access this feature from the right panel within any Bot (all users)
Design and UI Changes on Dynamic Insights*
Refreshed interface with improved usability
New feature: LLM selection options for customized insights
*Dynamic Insights is a Feature Flag (*Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain).
Folder Management for Insights
Create new folders directly via the "New" button
Select "New Folder" to organize your insights more efficiently
New Speech-to-Text Models (Feature Flag*)
Integration of Azure's advanced 4o-transcribe and 4o-mini-transcribe models
Access and configure these models via your profile settings (top-right corner)
*Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
Voice Input Overlay Optimization
Completely redesigned Speech-to-Text interface
More intuitive and responsive user experience
Multi-Select Connect/Disconnect for Data Room Sources
Use multi-select functionality to connect/disconnect multiple insights simultaneously
Currently available for Insights only (Files support coming soon)
Support for Image and Audio Files via SharePoint Connection
Consistent file handling across all upload methods
Seamless integration with existing SharePoint connections
Workflow Scheduling for All User Roles (*Feature Flag)
Access via "Set up workflow" in the Data Room
Click the three dots next to the workflow name to find "Workflow Scheduling"
*Feature Flags must be activated by your admin in coordination with the Key Account Manager from Blockbrain.
Improved Error Messaging with Fallback LLM
System now uses a fallback LLM to generate helpful error messages
Provides meaningful guidance when the primary LLM encounters processing issues
Automatic Full-Screen View for Long Chat Entries
System automatically expands to full-screen when entries exceed a certain length
Provides better readability for extensive conversations
Product Update – May 26th, 2025
Unified Navigation
The main navigation has been redesigned for a cleaner, more intuitive experience.
The Dashboard is now called Bots.
Knowledge Management and Admin tabs now feature drop-down menus for easier access.
As before, Consumers have access to Bots and Knowledge Management. Builders additionally see Workshop, while Admins have access to the Admin tab.
Seamless Session Continuation
Easily resume your work in the last data room you used when you reopen the application after closing the window.
Enable this feature once in Account Settings by selecting “Start on last data room”.
OneNote Access via SharePoint (Admin only)
Admins can now manage OneNote files directly via SharePoint integration for all users.
Web Component UI in German
The web component interface—e.g. used for integrating knowledge bots into SharePoint or websites—is now fully available in German.
Admins can switch the interface language between English and German.
Folder Upload
Upload entire folders (not just individual files) into a database source under Knowledge Management.
Enhanced “Save Insight to Knowledge Base”
Save insights more efficiently with a modernized workflow.
Select one or multiple Knowledge Bases and save insights into one or more folders within a Knowledge Base.
Support for 80+ File Types
Upload and process over 80 different file types, including VTT files.
Export Chat Messages & Insights as TXT
In addition to PDF and Word, you can now export chat messages and insights as plain text (TXT) files.
Use the multi-select option (three dots) under the prompt result to download as TXT.
Save Custom Prompts as Agents
Save your custom prompts as reusable agents.
Use the “Save as Agent” option (three dots) under the prompt result.
Prompt instructions are automatically included; just add a Category and Title (Description is optional).
New LLMs: Claude 4
Claude Sonnet 4 is now available on Vertex AI servers in the EU (like Claude Sonnet 3.7).
Claude Opus 4 is available exclusively on US servers for now.
For more details, visit: Claude 4 Information
Activation of Claude Sonnet 4 requires admin access and consultation with your Key Account Manager.
Redesigned Toast Messages
Toast notifications now have titles and a consistent look for clearer feedback.
Improved error messaging for system load failures (Error Level 0).
Improved Insights and Search
Enhanced prompts for Dynamic Insight and Full-text Search for better results.
Consistent Agent Responses
Agents now deliver more reliable and consistent responses, matching the quality of standard prompts.
Retry Failed Workflows
Easily retry any failed workflows directly from the interface.
Smarter AI-Generated Insight Titles
AI now generates more accurate and context-aware titles for insights.
Refined Sign-Out Modal for Connected Services
The sign-out process for integrations (e.g., OneDrive) is now clearer and more user-friendly.
Updated File Upload Instructions
File upload instructions are now clearer.
The maximum number of files applies per upload, not in total. Wording has been improved for better understanding.
Knowledge Bots – Product Update – May 9th, 2025
1. New Features & Enhancements
We have introduced a range of powerful new features to make working with Knowledge Bots even more efficient and flexible:
Download Chat Messages as PDF/DOCX Individual chat messages or results can now be exported directly as PDF or DOCX files—ideal for sharing and archiving key insights.
Enhanced SharePoint Integration Improved document visibility and optimized database source connectivity enable seamless collaboration.
Azure GPT-4o-Mini-TTS Voice Output Frontend support for voice output makes the platform more accessible and interactive.
Grok Integration Integration with Grok provides even more advanced AI options for demanding knowledge work.
New File Upload The file upload process has been completely redesigned for greater reliability and user-friendliness.
Webcomponent Bots: Notification & Retry for Crawling Errors Users are now notified of failed crawling attempts and can easily retry, ensuring data completeness.
2. Maintenance & UI Optimizations
Improved Dynamic Insight Usability Clearer labels, better hover states, and activated action buttons for a more intuitive user experience.
Updated Canvas and Dialogue Layouts Modernized interfaces and optimized error handling.
Audio Recording & AI-Powered Title Generation Insights can now be recorded via audio and automatically titled using AI.
Improved display and layouts in Audit Trail and LLM overview
Optimized multi-selection (checkboxes instead of radio buttons)
Knowledge Bots - Product Update - March 31st 2025
1. Enhanced Media Capabilities
We've expanded how you can capture and share knowledge:
Audio File Upload and Conversion - Upload audio files with automatic WAV conversion for transcription
Download Insights as PDF/WORD - Export insights in multiple document formats with preserved formatting
Increased File Size Upload Limits - Support for larger files in data rooms
Updated Static Insight Creation - Redesigned interface for intuitive insight creation (beta - feature flag only)
2. User Experience Improvements
We've enhanced platform interactions:
Dynamic Insights - Automatically update insights as new information becomes available (beta - feature flag only)
Token-based Presets - Optimize queries with predefined output lengths (beta - feature flag only)
Improved Error Handling - Automatic removal of error paths on page refresh
Knowledge Bots - Product Update - March 17th 2025
1. Expanded AI Model Portfolio - More Power, More Options!
We've significantly expanded our AI model lineup to give clients unprecedented choice and performance:
Google Gemini Flash 2 (US & EU regions) - Google's latest offering with impressive reasoning capabilities
GPT 4.5 integration - OpenAI's newest model with enhanced contextual understanding
Claude 3.7 Sonnet (Anthropic & Vertex) - Known for nuanced responses and excellent instruction-following
O1 platform feature (now out of Alpha) - Our proprietary model is now production-ready after extensive testing
2. Enhanced Search & Workflow Features - Streamlined Productivity!
We've reimagined how users interact with our platform:
Model switching in datarooms - Switch between AI models on-the-fly without disrupting workflow
Upgraded Insight Search with Global Search technology - Find exactly what you need, when you need it
Workflow access for all roles - Democratizing automation across organizations
Image Generation functionality - Expanding beyond text to visual content creation
Web Research Pro - Transform research capabilities with advanced real-time information gathering
3. User Management Improvements - Simplified Administration!
We've made administration more flexible and user-friendly:
Extended Temporary Role Switch for Admins/Builders - Test experiences from different perspectives
Automatic "General Access Group" assignment - Streamlined onboarding for new users
Enhanced Azure Group sync with pagination - Better handling of large organizational structures
4. Platform Experience Updates - Polished Interactions!
We've refined countless details to create a more cohesive experience:
Improved file upload consistency across the platform
Enhanced third-party integrations with Google Drive, OneDrive, and SharePoint
Live reasoning stream for Claude 3.7 and DeepSeek R1 - Watch the AI's thinking process in real-time
New hover effects with Dark Mode support - Subtle but important visual improvements
Knowledge Bots - Product Update - February 25th 2025
New EU-compliant models: OpenAI o1 and o3-mini are now available
These are still beta LLMs that are a) very expensive and b) have limited functionality - NOT recommendable to use for long workflows or web searches yet.
Document Chunk Highlighting: Helps users easily locate relevant sources in original documents
UI Improvement: LLM host region information is now visible in Bot Settings
Upload: ALL text-based formats are now allowed
Knowledge Bots - Product Update - 22 January 2025
Web Search Enhancement: Added Web Search Toggle to Workflow Interface
Knowledge Base Improvements: Implemented Email Service for File Upload
Speech-to-Text: Extended support to 50 minutes
SharePoint Integration: Updated Resync Permission functionality
Database Sources: Enabled .html file uploads to database sources
User Interface Enhancements: Added new "All Database Sources" tab in Data Management, management with edit icon for descriptions
Knowledge Bots - Product Update - 06 January 2025
1. Global Search
New full-text search functionality across the entire platform
Enables quick and efficient information retrieval across all areas
2. New AI Models & Compliance
New LLMs from Vertex AI available in Admin tab
Claude 3.5 SONNET V2 now EU-hosted
GDPR compliant
High-performance model for European users
3. Email Integration
Connect Gmail and Outlook accounts
Search through email inboxes
Direct integration of selected emails as insights into ongoing conversations
4. Workshop Permissions
New User Roles in Workshop:
Editor: Full editing rights
User-Editor: Can add and edit bots
User: Usage rights only
5. Dashboard Improvements
New "All" filter for bots
Simplified view of all bots
No longer limited to Retriever or Nexus selection
6. Web Search Agent
Activation of web search functionality
Enables web research during queries
Knowledge Bots - Product Update - 04 November 2024
Feature: Global Search
Description: The new Global Search feature allows users to search across the entire platform for relevant information quickly and efficiently.
Key Benefits:
Unified Search: Perform searches across all data sources and modules within the platform.
Improved Efficiency: Reduces time spent looking for specific information by providing a centralized search function.
Enhanced User Experience: Provides a seamless and intuitive search experience, making it easier to find and access necessary data.
Feature: Refine Email Newsletter and Knowledge Contribution Loop
Description: Users can now connect their email inbox to the Knowledge Management system, allowing them to pull specific emails as insights directly into the platform.
Key Benefits:
Email Integration: Seamlessly integrate your email account to import important emails into the Knowledge Management system.
Insight Extraction: Convert relevant emails into actionable insights, enhancing the knowledge base.
Streamlined Workflow: Simplifies the process of managing and utilizing email content within the platform.
Knowledge Bot - Product Update - 25 October 2024
Rework the Workshop Bot Access Control: You can now share your Bots with different access rights.
Ability to Switch Language to German on entire Platform
Human-in-the-Loop Mode for Automation Workflow: Workflows may now be checked step-by-step.
Improved UI/UX for Sharing Modal
Ability to update Builder-Role in Dashboard
Real-time Synchronization of Pinned Message Edits with Chat Timeline
Add "Created At" and "Modified At" Timestamps to User and Bot Messages
Share Selected Agent in Shared Bot/Add to New User's Dashboard - View Only, No Edit Permissions: Allows users to share a specific agent within a shared bot environment, automatically adds the shared agent to new users' dashboards for immediate access, and restricts new users to view-only permissions to maintain bot configuration integrity.
Knowledge Bot - Product Update - 09 September 2024
Merging Retriever and Nexus: Unified functionality and UX for improved efficiency.
User Analytics Tab in Admin Dashboard: New analytics features for better user insights.
More LLM Models from Vertex AI (Mistral and Llama): Expanded model options for better performance.
Integrate GPT-4o Mini from OpenAI and Azure: Enhanced AI capabilities.
Agent Prompt Library: Introduction of a centralized Agent Prompt Library (Agent Templates), allowing reuse of agents across different bots and data rooms.
Add Speech to Text Icon in Agent Library: Enhanced accessibility.
New Folder and Subfolder Structure for Document Management: Implementation of an improved structure for more efficient document organization.
Bulk Image Delete Function: Introduction of a feature to delete multiple images simultaneously in document management.
Integration of User Feedback: Continuous improvements based on feedback from our users.
Refined Design of Split Layout: Optimization of the display for pinned insights, improving user-friendliness.
Improvements in Reference Retrieval and Data Management: Overhaul of the user interface for more intuitive handling of references and data.
Knowledge Bot - Product Update - 05 August 2024
Shareable Knowledgebase Folder: New feature allowing users to save insights and messages into a shareable folder, complete with an accessible audit trail.
Bot Creation Flow and Settings Redesign: Streamlined process for creating bots and managing their settings, enhancing user experience.
Date Addition to Knowledgebase Contributions: Implementation of date stamps for messages and insights added to the knowledge base, improving traceability.
Dataroom-Specific Workflow Auto Trigger: Introduction of customizable workflow auto-trigger settings for each individual dataroom.
Knowledge Bot - Product Update - 26 July 2024
Excel files (xls, xlsx) for document processing: Enables users to upload and process Excel files directly within the platform.
New LLM Model: Launched the new LLM model, Gemma 2, which offers advanced capabilities (note: it is not optimized for speed).
Download option: Allows users to download files that have been uploaded to Nexus.
Admin Tenant Dashboard for LLM Models and Embedding Models: Introduces a dashboard for administrators to manage LLM and embedding models.
Agent for part of message in Nexus Bot | Scroll AI Context Menu: Allows users to select agents for specific parts of a message within the Nexus Bot and scroll through the AI context menu.
User Initials Avatar: Adds user initials as avatars for better identification.
Dragdrop and Agent Import to Workflow: Enable drag-and-drop functionality and agent import in workflows.
Default agents to be set by user preferences instead of a bot-level config: Allows users to set default agents based on their preferences rather than a bot-level configuration.
Knowledge Bot - Product Update - 08 July 2024
Release Notes (08 July 2024)
New Folder and Subfolder Structure for Document Management: Implementation of an improved structure for more efficient document organization.
Bulk Image Delete Function: Introduction of a feature to delete multiple images simultaneously in document management.
Integration of User Feedback: Continuous improvements based on feedback from our users.
Refined Design of Split Layout: Optimization of the display for pinned insights, improving user-friendliness.
Improvements in Reference Retrieval and Data Management: Overhaul of the user interface for more intuitive handling of references and data.
Knowledge Bot - Product Update - 01 July 2024
Release Notes (01 July 2024)
Renamed 'Folders' to 'Data Rooms' to reflecting the ability for user to connect different data and context .
Ability to pin an insight across different 'Data Rooms'.
Refined design of system messages.
Refined design of the right side bar in the Knowledge Bot Nexus.
Bot Settings
File Management (Context/Data that is connected to the respective data room)
Other Settings (e.g. Language of the bot)
Added detailed descriptions and parameters to benchmark different AI models
Model descriptions (Info about which model is suitable for which tasks)
Answer Quality (General intelligence and reasoning skills)
Performance (Response speed)
Cost Efficiency (Cost per input and output tokens)
Added the latest LLM models that are hosted on European servers.
GPT-4o (Microsoft Azure)
Claude 3.5 Sonnet (Vertex AI)
Gemini 1.5 Flash (Vertex AI)
Knowledge Bot - Product Update - 23 May 2024
Release Notes (23 May 2024)
Redesigned and refined the "document management" section to give users a better overview of their folders, files, and insights
Added Microsoft SharePoint Integration allowing administrators to sync and update documents seamlessly. Admins can selectively share folders from SharePoint that can be used as a knowledge base.
Implemented "Pinnable Insight Feature" which helps users to create and iterate on content/documents much more seamlessly. Users can now pin important messages and insights to the left side for easy reference while iterating on the content on the timeline to the right side.
Added the best performing LLM (AI) models on the market to date for users to select:
GPT-4o (Omni)
GPT-4 Turbo (Vision)
Claude 3 Opus
Claude 3 Sonnet
Claude 3 Haiku
Gemini 1.5 Pro (Vision)
Gemini 1.0 Pro (Vision)
Knowledge Bot - Product Update - 14 March 2024
Release Notes (14 March 2024)
Import textual & image content from webpages to build a knowledge base
Advanced document references that explains which information was used for answer generation
Advanced bot sharing option to restrict access to specific users in the organization
Multi-language support (Nexus)
Option to include chat history as a reference when sharing a note with another user (Nexus)
Last updated