Skip to main content

Uniphore Help Center Portal

2026.07 Real-Time Guidance Agent

Release Date: 24 August 2026

The Real-Time Guidance Agent 2026.07 Release introduces the following features.

New Features

There are no new features added in this release.

Enhancements
Session Management and Authentication Enhancements in AI Agent Assist

With this release, AI Agent Assist includes enhanced session management and authentication capabilities to improve security, reliability, and the overall Agent experience. Previously, Agents could experience unexpected session termination, '401 Unauthorized' errors, interrupted transcripts or events during active sessions due to authentication token expiration, and failed logout attempts.

Agents now receive notifications before inactive sessions expire, and sessions are automatically terminated after the configured timeout when the browser is closed.

The application also improves the logout and re-authentication experience by performing enhanced session cleanup and validation. Authentication tokens are automatically refreshed to help prevent interruptions during active or temporarily idle sessions. If token renewal fails, the application prompts the Agent to re-authenticate.

For more details, refer to Logging In to the AI Agent.

Constraints and Limitations

The following constraints and limitations exist for this release:

Item

Description

Workaround

1

Currently, the Manager Dashboard only supports live monitoring for up to 200 Agents simultaneously per Manager.

When using the Manager Dashboard for live monitoring of call sessions in your account, do not map more than 200 Agents to a Manager. Contact Uniphore Support for additional information and options.

2

When using the Auto-Submit option with the GenAI Summary, if the Trigger is set to 'End of Current Call', the X-Platform Conversation Summary variable for the current call returns an empty value.

Contact Uniphore Support for additional information and options.

3

When creating an Alert for Agents, you can currently map it to a Service Flow; however, the selected Flow will be executed only at the end of the call.

Execution of a Service Flow that is mapped to an Alert will be possible only in a future release.

Bug Fixes

The following issues were fixed in this release:

Item

Description

1

French-language intents were not being trained successfully in the production environment. This issue has been resolved, and French-language intents can now be trained successfully for French-language tenants.

2

In the Knowledge Assist panel, responses to manual questions were not populated when a transcript was uploaded and remained in the Fetching state. This issue has been resolved, and responses to manual questions now populate successfully for both transcript and audio uploads.

3

When testing a summary template in Summary Builder, the test status could remain stuck in the Under Processing state even after processing was completed. This issue has been resolved, and the test status now automatically updates to Completed or Failed when processing finishes.

4

There were differences between summary display, manual submission, and auto-submission metrics in the GenAI Summary Metrics dashboard, which could lead to an inaccurate interpretation of summary activity. This issue has been resolved, and the dashboard now provides improved visibility into summary generation, delivery, and submission events.

5

When the AI Agent Assist application was configured with a single-column layout, the application did not always open the panel selected as the Default Opening Panel. This issue has been resolved, and the configured panel now opens correctly when a call starts.