BA Insight release notes
The following features are new in this release of BA Insight. To see previous releases, see the archived releases topic.
BA Insight 2026 R1.0.0 release
SmartHub
This outlines all of the updates and enhancements in this release of SmartHub. For more information on upgrading, see Upgrading SmartHub.
The following features and enhancements have been added to SmartHub in this release:
System enhancements
| Feature | Description |
|---|---|
| Support for Agentic RAG |
SmartHub now supports configuring your environment with agentic RAG capabilities, allowing for improved search result retrieval based around user intent and accurate, context-aware answers through multi-step reasoning and hybrid retrieval. Instead of relying on a single search and response step, agentic RAG allows SmartHub to reason, retrieve, and act across multiple stages before generating a response. This release introduces configurable AI agents that orchestrate query processing, semantic search, Knowledge Graph traversal, and response generation in parallel. To configure agentic RAG in your environment, refer to the following topics:
Supports upgrade from SmartHub 7.2 to 7.3 by importing the legacy FederatorConfiguration.xml, allowing existing Conversational Search settings to be reused with some manual adjustments. Currently supports AzureOpenAI and OpenAI models. Other AI models can be used, but an API action must be created in Orchestrator.
|
| Support for Azure Cosmos DB for knowledge graph retrieval |
This release adds support for Azure Cosmos DB as a search engine. With Azure Cosmos DB, SmartHub can build a Knowledge Graph that captures how information is related, not just where keywords appear. This enables richer contextual search experiences and supports advanced use cases that rely on understanding relationships between data. For more information, see Setting up your Azure Cosmos search engine. |
| Support for Knowledge Graph entity extraction |
SmartHub has introduced new tuning stages to support knowledge graph entity extraction from user search queries. This stages utilizes large language models at query time to identify and extract meaningful entities and relationships from a user’s search query. You can configure domain-specific ontologies to control what entities are detected and how they are linked, improving search relevance, contextual reasoning, and downstream agent workflows For more information, see Utilizing entity extraction for search queries for OpenAI or Azure OpenAI. |
| Multi-topic search query enhancements |
SmartHub has introduced new tuning stages to support topic splitting for multi-topic search queries. This allows you to leverage OpenAI and Azure OpenAI resources automatically split a user's search query into sub-queries, processing them in parallel and intelligently synthesizing them into a single response, improving relevance and depth while respecting existing search limits and configuration settings. For more information, see Configure the Topic Splitter tuning stage for OpenAI or Azure OpenAI. |
| Support for downloading documents from conversational search responses |
SmartHub now has the ability to create a downloadable document when it is requested in conversational search based on the user's chat history. By default, the Orchestrator agent comes configured with the Download Required action, allowing your conversational search configuration to leverage your large langauge model to accurately respond to user's natural language request for a document in .pdf or .docx format. For more information, see Understanding pre-populated agent actions in the agents management documentation. |
| Analytics reports for conversational search feedback. |
SmartHub analytics reports now display the feedback results when a user specifies whether they liked or disliked a response that was returned by conversational search. The following metrics are collected:
For more information, see analytics reports. |
| User experience tuning stage enhancements |
User experience tuning stages have introduced a new Trigger Condition field, allowing you to specify a scripted condition to determine if a specific tuning stage should execute. For more information, see add a tuning stage. |
| Search menu enhancements |
The SmartHub search menu has been enhanced with the following setting:
For more information, see configure the Search Menu. |
| Credential storage enhancements |
You can now store your credentials and sensitive information in Azure Key Vault for retrieval by a SmartHub script tuning stage. This prevents sensitive information from being included or exposed in your SmartHub scripts. For more information, see the instructions for configuring Azure key vault storage in Configure your storage settings. |
| Spellchecker enhancements |
The following setting has been added to the SmartHub spellchecker function:
|
| Saved queries and query history enhancements |
The following setting has been added to the SmartHub saved queries feature:
|
| Microsoft O365 user profile provider enhancements | Microsoft O365 user profile providers can now retrieve user profile properties from SharePoint Online, Microsoft Graph, or both, depending on your configuration. For more information, see Configure user profile and picture providers. |
Fixed issues
The following issues have been fixed in this release:
-
If you applied a filter that did not exist to a search query, the search results would return all results, rather than no results. (1831672)
-
In some scenarios, the SmartHub analytics module was producing inconsistent results due to page sessions expiring. (1817622)
-
When configuring a search engine, switching from Regular to Experimental would cause an invalid type message to display in the Search Engine field.
-
In some scenarios, the filter button was absent from the conversational search page.
-
when applying changes to a container from the SmartHub administration portal, the changes took a long time or required an iisreset to be reflected.
-
When opening the conversational search settings page, console errors appeared.
-
When setting up an expertise page with the Entities based on related – Large pool model, date refiner range facets were not filtering the number of users returned.
-
In a Microsoft Teams work hub integration, clicking the "Go to Website" button would open a SmartHub page with the search bar missing.
-
When using a Microsoft Search Items search engine with external items, any range facets, such as date refiners, were not loaded.
-
Duplicate date range refiners were displaying in the refiners container.
-
The Add to Basket button was not changing color after a search result was added to the users basket.
-
If the sort direction in the SmartHub results page URL started with an uppercase, the content container for the sorting mechanism would fail.
-
When exporting results from the SmartHub results page, nothing was downloaded to the user's browser.
-
When SmartHub logging was set to debug mode, the logs were only showing warnings and errors.
-
The administration portal link from the profile menu on the SmartHub index page was redirecting to a 404 error page.
-
UI issues that affected the visual design and layout of SmartHubs refiners.
Connectivity Hub
This outlines all of the updates and enhancements in this release. For more information on upgrading, see How to Upgrade Connectivity Hub.
Connectivity Hub 4.4.0.1 release
The following features and updates are new in this release of Connectivity Hub:
System enhancements
There are no system enhancements in this release.
Fixed issues
The following issues have been fixed in this release:
-
Content reset jobs were failing with a "Could not load file or assembly" error. (1909754)
-
Security sync jobs were failing with a "Given key was not present in the dictionary" error for connections that were configured with a group mapping script that threw an exception. (1901808)
-
When Push incremental crawl mode was enabled for a content source, push crawls were failing or continually looping . (1856381)
-
Crawl jobs were failing when the URL to Item Resolver target specific setting was configured. (1892744)
-
In a multi-server environment, if identity provider authentication was configured on one server, and the load balancer URL pointed to a different server, the client secret was not be able to be decrypted, causing Connectivity Hub to be unusable.
-
When Push incremental crawl mode was enabled for a content source, push crawls were not deleting datastores from the Connectivity Hub database after they had been deleted from the source system.
-
When Push incremental crawl mode was enabled for a content source, push crawls that intended to delete folders containing items were not deleting the folder, only the items within it.
Connectivity Hub 4.4.0.0 release
The following features and updates are new in this release of Connectivity Hub:
System enhancements
| Feature | Description |
|---|---|
| Support for Amazon Q access control lists |
The Amazon Kendra target has been updated in this release to be compatible with Amazon Q. Connectivity Hub now pushes both SmartHub ACL custom metadata and native Amazon Kendra ACLs when indexing content. This allows Amazon Q to use the same Kendra index, with proper document-level security enforcement via IAM Identity Center, without impacting existing SmartHub integrations. For more information, see Configure Your Kendra Target. |
| Security performance and stability improvements |
This release adds improvements Connectivity Hub security, reducing the impact to environments with large or complex security datasets, strong security guarantees and compliance readiness, and incremental delivery. |
| Connectivity Hub administration portal authentication |
The Connectivity Hub administration portal now supports identity provider authentication rather than basic windows authentication, allowing users to log in through enterprise credentials and multi-factor authentication policies. |
Fixed issues
The following issues have been fixed in this release:
-
Resolved an authentication exception when using SQL auth after upgrading Connectivity Hub. (1893108)
-
Resolved an issue where running an incremental crawl after a full crawl produces an error. (1850111)
-
For some users, calls to the advanced security module were taking excessively long to complete. (1851107/1850946/1841066)
-
If user membership was removed from a group, the membership removal was not reflected after the subsequent security sync.
-
Due to issues related to performance, the security mapping page was not displaying the number or child users/groups. As a result, the UI controls to access this page were removed.
-
In some scenarios, after upgrading to Connectivity Hub 4.3, a "The ConnectionString property has not been initialized" error displayed when attempting to connect to the database.
-
After upgrading to Connectivity Hub 4.3, crawling a content source that existed pre-upgrade caused the crawl to fail with an Exception has been thrown by the target of an invocation - Object Reference not set to an instance of an object error. (1849454)
-
If the source ID for the user mappings of a connector had the same target directory account in Azure Active Directory, calls to the advanced security module would fail. (1837531)
-
In some scenarios, after upgrading to Connectivity Hub 4.x, public groups that were assigned to the ACL were not being found. (1835843)
-
After cloning a content source, the target specific settings were not maintained from the original content source. (1812853)
-
In some scenarios, after upgrading to Connectivity Hub 4.2, and extra log purge jobs was incorrectly created. (1799379)
-
Due to a limitation on the request size for Azure App Service on IIS, large files were failing to be indexed. (1795573)
-
In some scenarios, an Object Reference not set message appeared in the top ribbon of the Connectivity Hub administration portal. (1791308)
-
During enumeration, if a connector returned duplicate items, the enumeration would fail, resulting in discrepancies in the search results. (1786374/1774157)
-
In some scenarios, datastore type load tasks were failing with an Exception SyncingException error.
-
In some scenarios, when Connectivity Hub was configured with Active Directory as the target directory, calls to the API were failing to identify and map the group members.
-
If a log purge task was created with the Delete Unused acl groups setting enabled, the task would fail.
-
After creating a cache database using an SQL account with proper permissions, attempting to reassign the db_owner permission to another SQL account was producing an error and the database would not be reassigned.
AutoClassifier
This outlines all of the updates and enhancements in this release. For more information on upgrading, see How to Upgrade AutoClassifier.
The following features and updates are new in this release of AutoClassifier:
System Enhancements
| Feature | Description |
|---|---|
| Support for Knowledge Graph entity extraction |
New Knowledge Graph entity extraction components, powered by OpenAI and Azure OpenAI, have been added in this release. These components extract structured entities and relationships from indexed content, enabling your SmartHub to use graph-based reasoning during conversational search. This allows SmartHub to move beyond keyword and similarity search by leveraging relationships between entities when answering questions. To configure Knowledge Graph entity extraction in your environment, refer to the following topics:
|
| Document chunking enhancements |
The document chunker component has been enhanced to only return the most relevant chunk in user search results. This prevents the chunk component from returning repeat information for a search query. The document chunker component will now include the chunk ID in the output properties. You can append additional properties to the first chunk using the Properties for First Chunk setting. For example, you can append the entire full text of the document in the first chunk by appending the body property. For more information, see Configure the document chunker component. |
| Table of contents extraction |
You can extract section names from the a document's table of contents with the Section Annotator component. This extracted section name can then be applied to the metadata of a document chunk, improving the relevancy of the search results, allowing end users to accurately filter their search results by the section name, and improving provenance metadata to explain where a document chunk came from. For more information see Extracting the table of contents for metadata appendation. |
| AutoClassifier administration portal authentication |
The AutoClassifier administration portal now supports identity provider authentication rather than basic windows authentication, allowing users to log in through enterprise credentials and multi-factor authentication policies. For more information, see Filters. |
| Ability to force clear taxonomy synchronization |
You can now clear your taxonomies table from the configuration page using the Force Clear Taxonomies button. Once cleared, you will need to synchronize configuration to populate your tables. For more information, see Using the Add-in configuration page. |
| Deprecated features for SharePoint add-in |
As of April 2nd, 2026, Microsoft is ending support for SharePoint add-ins in Microsoft 365. As a result, AutoClassifier is deprecating support for the high-trust certificate security model in the SharePoint add-in. |
| AutoClassifier license improvements |
The logging mechanisms for AutoClassifier license validation have been improved in this release. The AutoClassifier logs will now capture the license validation status and the reason for any failures should they occur. Additionally, the AutoClassifier UI will now display improved error messages for license-related issues. |
Fixed issues
-
If a SharePoint Online document library was specified as Do Not Index, the SharePoint Online add-in would not annotate the library. (1840408)
-
The contextual metadata merger component was failing is no content was sent from Connectivity Hub to AutoClassifier. (1793654)
-
When creating or saving changes to an adapter source, a null reference exception was returned in the audit log.
-
The AutoClassifier license expiration message was displaying on the expiration date, rather than one day after the license expired.
-
Changes to the sequence of components in a pipeline were not recorded in the audit log.
-
When annotating columns from SharePoint Online, if a script was used to update fields, but not all of the fields were included in the SharePoint Fields to Annotate configuration field, the annotation process would fail.
-
If the Restful API Caller component was added to a pipeline directly after the Document Chunker component, the chunked child items were not present in the output.
Smart Previews
This page outlines all of the updates and enhancements in this release. For more information on upgrading, see How to Upgrade Smart Previews for SharePoint O365, SharePoint 2016/19, or SmartHub.
System Enhancements
| Feature | Description |
|---|---|
| Support for modern Microsoft Office documents | Smart Previews now support the latest version of OpenXML to generate previews for modern Microsoft Office documents. |
Fixed Issues
-
Added support for CSV and XLSM file types. (1830784)
-
When registering fast proxy, the registerproxy.bat file was failing when run as an administrator or from the command prompt. (1779672)
-
The fast proxy service was not starting in an environment with an Azure database and Active Directory Service Principal authentication selected.
Breaking changes
The following are breaking changes that may affect your integration. Please ensure that you carefully review this list before upgrading your environment:
SmartHub
-
When upgrading to SmartHub 7.0 and later, you must update any custom pages to preserve your existing CSS or upgrade to the out-of-the-box SmartHub 7.0 page layouts. For more information on this update, see Update or preserve your custom CSS in the upgrade documentation.
-
The DomContentLoaded event has been removed in SH 6.2. You will need to replace the event with the afterInitialization event. For more information, see UI events.
Connectivity Hub
-
Due to the new crawler logic, the number of sync threads is now applied as a content specific setting. The number of sync threads that were specified will now be applied to every job that is concurrently running for the same target.
-
When upgrading from versions prior to Connectivity Hub 4.0 or later to integrate with AutoClassifier 7.0 or later, the Service URL that is used to communicate with AutoClassifier must be updated. For more information, see Using AutoClassifier components to enhance metadata.
-
Layer 4 load balancers (NLB) are no longer supported if they must preserve the client IP. Layer 7 load balancers (ALB) must be used instead.
-
The default installation folder is now <program files>\Upland BA Insight\ConnectivityHub\. Please follow the install and upgrade instructions to ensure a seamless transition when upgrading to Connectivity Hub 4.0 and later.