Release Notes: SearchUnify Q3 '25

We’re excited to announce the Q3 ’25 release of SearchUnify, featuring smarter analytics, stronger privacy controls, and an enhanced search experience. With new content sources, refined hybrid search, and advanced reporting, this release empowers admins and users to drive more efficient and smarter user experiences.

Key Highlights

The key highlights of the Q3 ’25 release include:

  • New content sources: Absorb LMS and Amazon S3

  • Stronger masking of personal information in search results

  • Support for the range filter on number values

  • Improved and context sensitive Hybrid Search

  • New SearchUnifyGPT™ Engagement Report

  • Detailed Admin logs for Content Sources and Search Clients

  • Support for up to four alternative responses in Response Assist

Content Sources

New Content Source: Absorb LMS

Absorb LMS is a learning management system designed to streamline employee training and development. Absorb LMS enhances learning experiences and improves organizational performance. Admins can now integrate Absorb LMS with SearchUnify to index learning materials.

More information: Use Absorb LMS as a Content Source

New Content Source: Amazon S3

SearchUnify now supports Amazon Simple Storage Service (S3) as a Content Source. Amazon S3 is a robust and widely-used cloud storage service that allows companies to store and retrieve large volumes of data. With this integration, the SearchUnify admins can index content from their S3 buckets and make the data searchable for end-users.

More information: Use Amazon S3 as a Content Source

Stronger Data Protection in Search Results

Using a mix of regular expressions, natural language processing, and other techniques, SearchUnify will now mask personal information from the search results.

For example, instead of “The case owner is Kevin William” the search users will see “The case owner is <PERSON>.” To view the name, search users will have to open the document. If they have the right set of content source permissions, then they can read a person’s name.

The following information is masked:

  • Names

  • Phone numbers

  • Addresses (or locations)

  • URLs

  • Pincodes

More information: Preview Search Results

Search Clients

View Results Within Specified Value Range

A new filter type, Range Slider, is introduced for content fields that are used to store numbers and date-time values. For example, now search users can filter results by topic IDs that fall in the range 192732 -196677.

Fig. A snapshot of Range Slider on a search results page.

The Range Slider works on number and date-time fields. This means that SearchUnify admins can configure it on fields, such as Case Creation Date and Case IDs.

More information: Configure Content Sources Linked to a Search Client

View Status of Recrawl on Merge Results

When results are merged, the associated Content Source is crawled again. The status of the crawling process is displayed directly on the Merge Results tab. The following statuses are available:

  • No Change: The Merge Results tab remains unchanged in two cases:

    • The feature has not been configured.

    • The feature was configured but has been disabled.

  • Orange Dot: An orange dot and a blue border appear on the Merge Results tab. This happens when:

    • Merge Results has just been configured.

    • Recrawling is in progress.

  • Green Tick: A green tick and a blue border appears on the Merge Results tab. They indicate that:

    • The recrawl was successful.

    • The Merge Results tab is active.

  • Red Dot. A red dot appears when the recrawl has failed. In this case, admins can open the Merge Results tab for editing and save the configurations. This will automatically trigger a new crawl.

Fig. A snapshot of the Merge Results button showing all the possible crawling states.

More information: Merge Multiple Versions of a Doc into One Result

Smarter, Sharper Hybrid Search

Hybrid Search has been made more intelligent. Previously the algorithms in Hybrid Search activated keyword search on exact match queries, such as case IDs, and vector search on broad queries, such as the “cases of type X” but the weightage could not be customized. Now, the approach is different.

Both vector search and keyword search are used simultaneously to provide contextually relevant search results. An advantage of this approach is that if Field Boosting and Hybrid Search are active simultaneously, then giving a higher weightage to keyword matches delivers better results. The weightage of vector search to keyword search can be tweaked by the SearchUnify team on customer request.

Fig. A snapshot of the improved Hybrid Search.

Granular Vectorization, which is search client specific, is supported. In the backend, technologies such as embedded type support, partial vectorization at the level of search client, and up to nine embedding models are now supported.

Hybrid Search is turned off by default. To turn it on, write to your CSM. Before turning on Hybrid Search, ensure that all the required fields are checked in the Searchable column. When Hybrid Search is turned on, users cannot add or remove searchable fields. The entire column is disabled.

Fig. The Searchable column cannot be edited because Hybrid Search is on.

Searchable fields cannot be edited or removed from any content source linked with the search client where Hybrid Search is on.

Re-Ranking is also incompatible with Hybrid Search. You can use either Re-Ranking or Hybrid Search, but not both simultaneously.

Searchable fields can be edited after disabling Hybrid Search. Disabling and enabling Hybrid Search, vectorization takes place again. Until the re-vectorization is complete, the search will function on the old vectors.

Fig. A warning about re-vectorization when Hybrid Search is disabled.

More information: Neural and Hybrid Search

Search Analytics

View User Interaction with SearchUnifyGPT™ Engagement Report

A new report, SearchUnifyGPT™ Engagement, has been introduced in Analytics > Search Analytics > Overview. This report displays statistics on how search users interact with the SearchUnifyGPT™-generated answers. Using this report, SearchUnify admins can see how many users:

  • Read the SearchUnifyGPT™-generated answers

  • Click on the citations in the SearchUnifyGPT™-generated answers

  • Copy the SearchUnifyGPT™-generated answers

Fig. The SearchUnifyGPT™ Engagement report in the Analytics module.

More information: SearchUnifyGPT™ Engagement

Security

Streamlined Admin Logs for Content Sources

Admin Logs now capture only changes made under the Content Sources tab, eliminating unrelated user activity to provide cleaner, more relevant records.

Changes are now logged for actions like adding, editing, or deleting content sources, updating authentication or crawl frequency, modifying rules or annotations, managing crawl actions, and handling document operations.

Users can now click View Details to see changes, organized by tabs such as Authentication, Configuration, Frequency, Rules, Content Annotation, and Crawling Cases.

Fig. A snapshot of the Admin Logs for changes in the Content Sources tab.

More information: View Instance Manager Activity in Admin Logs

Streamlined Admin Logs for Search Clients

All changes under the Search Clients tab are now consolidated for easier tracking. The new logs are more detailed yet cleaner.

The changes made in the following tabs are captured in the logs:

Users can now click See More to view changes, organized by tabs.

Fig. A snapshot of the Admin Logs for changes in the Search Clients tab.

More information: View Instance Manager Activity in Admin Logs

Agent Helper

Feedback Analysis: Filter by Feature

Admins can now filter the Feedback Analysis report by feature to focus on the most relevant insights.

Available filters: All, Case Summary, Case Timeline, Response Assist

Fig: Snapshot of the filter added in the Feedback Analysis report.

Response Assist: Regenerate and Compare

Agents can generate up to four alternative responses using Regenerate. All versions are preserved for side-by-side review, making it faster to select the best response.

Case Timeline (Zendesk): View Internal Comments

The Case Timeline in Agent Helper now displays Internal Comments for Zendesk tickets alongside status changes, public comments, and escalations for deeper context and better visibility.

Note: This capability was previously available for Salesforce; it is now supported for Zendesk as well.

Slack Case Swarming (Zendesk): Upgraded Message Template

When an agent initiates Case Swarming via Slack, the shared message now includes a richer template with:

  • A hyperlinked Case ID

  • Case Summary

  • Preliminary Response draft

This improvement, earlier available for Salesforce, is now extended to Zendesk to ensure a consistent, expert-friendly handoff and faster, higher-quality feedback.

Other Updates

  • Implicit Case Deflection now available on Salesforce Communities: In the first quarterly release of this year, True Implicit Case Deflection was introduced for web-based search clients. It allowed the admins to implement a survey questionnaire on the search page where the users could share text feedback about their search experience. The admins could, through the questionnaire, find out:

    • if the users got what they were looking for

    • if the users would have contacted support if they hadn’t found the desired information

    • relevancy and effectiveness of the search results

    With True Implicit Case Deflection, the SearchUnify admins can better understand the rate of implicit deflection, i.e. users resolving their issues without creating a support case.

  • Clearer Instructions in README Files: The instructions for configuring a Search Client are included in the README file, which can be accessed by downloading the Search Client as a ZIP file. The README has been updated to improve consistency, logical flow, and readability. The UID and CDN endpoints are now clearly defined, making it easier for users to configure the Search Client by following the instructions.

    Fig. A comparison of the old (left) and new (right) readme files for the Salesforce search client.

  • Salesforce Communities Search Client is WCAG compatible: The search client for Salesforce Communities has been tested on the parameters of inclusivity and are now WCAG 2.2 standards compatible.

Known Issues