WhosOn 10 Release Notes
WhosOn 10 improves performance over previous versions, increases security compliance, requires fewer resources and disk space and reduces running costs.
Server Improvements
The WhosOn 10 server version numbering has changed. Previous versions used 'year' as the version (eg: WhosOn 2016, WhosOn 2019).
Core Server Changes
All assemblies now use .NET Framework 4.8 and are 64bit only. This simplifies installation and reduces setup size significantly.
Core TCP (websocket/http server) components used by the WhosOn Server, Gateway and Client have been updated to the latest versions with security improvements and support for TLS 1.3 and IPV6. All other core components have been updated to latest.
Simplified Installation
When installing the WhosOn 10 Server on a clean Windows Server computer, the setup will now install IIS and automatically configure IIS ARR rules. No changes are required to WhosOn ports.
Database Independence
The WhosOn Server database can now use SQLite, MySQL and PostgreSQL in addition to SQL Server. The benefits of this change mean reduced running costs since SQLite, MySQL and PostgreSQL are free to use for any DB size. The built-in SQLite database requires no external database server and provides good performance for installations with less than 10 users and less than 100 chats per day.
Simplification Of The Database
The WhosOn Database has been simplified and flattened. There are no stored procedures or functions. Separate tables for 'header' information have been removed. This significantly improves performance and simplifies reporting.
Database Archiving
Database archiving is now enabled by default. When the WhosOn Database is created, a separate archive database is also created (this uses the same database type as the main database - but can be changed later). Each site can have its own archiving settings - which will override the default.
Archiving can now use any of the supported database types (SQLite, MySQL, PostgreSQL and SQL Server) plus MongoDB (previously WhosOn could only archive to an Azure CosmosDB). The MongoDB can be local or any cloud MongoDB compatible service (such as Cosmos).
The Archive database has a reduced tables set, with flattened tables for Chats, Chat Lines, Chat Surveys & Visits - these are further flattened from the main database tables - and are designed for reporting.
Historical reporting can now use the archive database which reduces the load on the main database.
The default chat/visit retention days for the main DB have been reduced to 90 days.
Database Maintenance
The Database Maintenance can now run at multiple times of the day (eg: at 02:00, 06:00, 13:00 and 20:00). The run times are defined in the server settings. This splits up the deletion of data, so a single run will not cause database timeouts or usage overages.
Server Settings
Server wide settings are now stored in Json format in the 'ServerSettings.json' file.
MetaData
All site properties and user, user group, work period, skills, site images are now stored as Json in a LiteDB document database called 'MetaData.db'. XML is no longer used for any metadata/server settings. There is a 'Sites' and 'Users' table in the database, but this is only used for storing site/user names when reporting. No metadata is stored in the WhosOn database.
This allows easier reconfiguration of the database and transfer of metadata between servers - since the file can simply be copied.
The WhosOn Server is the only component that reads/writes metadata. This reduces possible errors, since the server becomes the single point of authority relating to metadata storage.
Changed ChatUID
The ChatUID is now a Unix-style object id (eg: 65b3b5f6c937df03142c7e51). These are fast to create and are sorted by datetime. They are unique (17 million per second before a possible collision) so the WhosOn server does not have to check if a new ChatUID already exists in the database.
Canned Responses (Auto-Responses)
Canned Responses are now referred to as 'Auto-Responses'. Auto-responses are added/updated/deleted via the Client interface. Auto-responses are no longer stored in the main database. They are stored in a document database called MetaDataArticleStore.db.
Knowledge Store Articles
Knowledge Store articles can now be created via the Client Interface. These are similar to Auto-Responses, but are designed to be used as context for AI based bot responses.
Document files (pdf, word, html, opendoc, markdown, html) can be imported. These will be converted to text and added. Large documents will be split into pages and saved as separate articles.
Articles are stored in a document database called MetaDataArticleStore.db in Json format.
Documents
User uploaded documents and visitor uploaded documents are now stored in the MetaData.db file. User documents are now uploaded via the client interface. The actual file contents are stored separately in a LiteDB file store called MetaDataFileStore.db. Uploaded files are no longer stored as physical files. This improves security.
Client Interface
The Client Interface now uses secure websockets only, with Json for commands and events. Therefore WhosOn 10 is not compatible with clients from previous versions.
The websocket interface is secure only (wss://) and uses a self-signed certificate if a trusted certificate has not been added to the server.
Chat Interface
The Chat Interface now uses secure websockets only, with Json for commands and events.
AI Integration
WhosOn 10 has AI features built-in as standard.
AI Settings
Each site now has properties for AI. The user can choose which AI service to use (currently ChatGPT or a locally installed OptimaGPT Server). Once an AI service type is added to a site then built-in bots, article embeddings, translation, sentiment scoring, summarization and classification can be enabled. A server-wide AI configuration can be created - all sites will use this unless a site-specific AI configuration has been created.
Built-In AI Bots
Bot users can be created. Bots will use the AI settings defined at the site or server level. The WhosOn Server handles AI bots internally - without the need for the separate bot service. The built-in AI bot can take and respond to chat requests. Bots will handle default context, knowledge context, conversation context, transfers and transcript requests automatically. Using bots with your own knowledge store now enables bots to be created in minutes. Users simply need to add auto-responses and articles to improve bot responses.
Built-In User To Bot Chat
When using AI bots, user can chat to the bot user via user-to-user chat. This is useful for testing bot responses and when users need answers to questions.
Auto-Responses & Article Embeddings
If an AI service has been added to the server or site, then embeddings will be added for any Auto-Response and Article added or updated. Embeddings enable semantic search of articles. The 'Articles' data will now become the local bot 'knowledge store'.
Chat Suggestions
If the Suggest Responses user property is enabled then WhosOn will automatically suggest responses to new visitor chat messages. If an AI service has been added, the suggestion will now use semantic search for Auto-Responses and Articles.
Bot Chat Monitoring (Agent Assist)
AI bots can now monitor other user's chats allowing for full user/bot hybrid chats. Users can request monitoring in the normal way. If no regular users are online or do not accept the monitor request, then the bot will accept the monitor request. The bot will then process the chat as if it was chatting to the visitor and send Whisper messages which will be a full bot response (using context from the Article search). The user can then choose to use the Whisper response for their own response (editing if needed).
Translation
Translation now uses the configured AI service instead of the Parker Software Translation web services. This provides improved translation from/to any language.
Chat Sentiment/Summarization
Chats are now summarized and sentiment scored using the configured AI service. This provides more accuracy and works without training and for any language.
Chat Classification
When an AI service is configured, chat tags can now be automatically generated (eg: 'Is this chat support, urgent support or sales'). Auto-generated tags will be added to any user specified chat tags.
OptimaGPT
For AI-powered features, WhosOn can use either OpenAI’s ChatGPT or Parker Software's OptimaGPT - an on-premises or private cloud-hosted AI server. OptimaGPT offers localized AI processing, ideal for companies needing to avoid external data transfer due to privacy regulations. OptimaGPT allows organizations in regulated sectors to deploy AI securely while meeting data protection and compliance needs.
For more details on OptimaGPT, you can visit the OptimaGPT website.
WhosOn Client
New Windows Client
A new version of the existing Windows Desktop Modern Client has been created that works with WhosOn 10. This has an updated look and feel and supports full editing of site/user properties (if the user login has sufficient rights) along with server settings, registration, database create etc. The new modern client has its own 'check for updates' option that provides a more seamless update process (this can be disabled server-side).
New Cross Platform Client
A new cross-platform client replaces the existing web client. It uses web assembly so will run on any platform. The new client can also be used to edit site/user properties (if the user login has sufficient rights).
No Service Manager Or Settings Portal
The Service Manager is now part of the new Client. There is no 'settings portal' in WhosOn 10. All configuration is performed using the WhosOn Client.
Chat Window
New Chat Window
A new chat window has been created that uses latest modern frameworks for improved performance and security compliance. The Chat Window is a .NET Core self-contained application that can run on Windows or Linux.
Upgrading From WhosOn 19
WhosOn 10 is not compatible with previous versions of the WhosOn database or settings. WhosOn 19 data and settings can be migrated to WhosOn 10 using the new WhosOn Client. WhosOn 10 can be installed on the same computer as WhosOn 19, however it is recommended to install WhosOn 10 on a clean machine and then migrate the database. If installing WhosOn 10 on the same machine, you should uninstall WhosOn 19 first (the database will not be removed), remove the 'chat' website from IIS and remove any ARR server farm & rules (since WhosOn 10 will setup new versions of these).
If you are using a version of the WhosOn Server prior to 19, you will need to upgrade this to the latest WhosOn 19 first.
Once WhosOn 10 is installed and the first-time setup completed, start the WhosOn Client and login with the System Administrator user.
Select Settings : Server Settings : Database tab. Click the Import WhosOn 2019 button.
You will need to enter the WhosOn 19 database connection string. This will be populated automatically when installing WhosOn 10 on the same machine as WhosOn 19 was installed. If installing on a different machine, you will need to obtain the WhosOn 19 database connection string. This can be found using the WhosOn 19 Service Manager application. Select Server Settings : Database. The connection string will be shown. Alternatively the connection string can be found in the Registry at
HKLM\SOFTWARE\WOW6432Node\ParkerSoft\WhosOn\7\Settings\DBConnection.NET
You may need to reenter a password manually in the connection string, since the registry value stores any password encrypted.
Click the Open button to open the WhosOn 19 database using the connection string provided.
All sites will be listed. Select the sites you want to import into WhosOn 10. Enable the Import Chat Data if you want to import previous chat records. You can adjust the starting data for chat data import in the Import Chat Data Starting From Date entry.
Click Start to begin.
(c) Parker Software 2024 https://www.WhosOn.com